Enhancing the performance of semantic segmentation models with multi-spectral images (RGB-IR) is crucial, particularly for low-light and adverse environments. While multi-modal fusion techniques aim to learn cross-modality features for generating fused images or engage in knowledge distillation, they often treat multi-modal and missing modality scenarios as separate challenges, which is not an optimal approach. To address this, a novel multi-modal fusion approach called Optically-Guided Pixel-level contrastive learning Network (OGP-Net) is proposed, which uses Distillation with Multi-View Contrastive (DMC) and Distillation for Uni-modal Re- tention (DUR) to maintain the correlation between modality-shared and modality-specific features. DMC aligns the uni-modal features by projecting the semantic information across modalities into a unified latent space, ensuring that the feature maps retain multi-modal representations. Pixel-level multi-view contrastive learning is introduced to enable modality-invariant representation learning. To retain modality-specific information, DUR is proposed, which distills detailed textures from RGB images into the optical branch of OGP-Net. Additionally, the Gated Spectral Unit (GSU) is integrated into the framework to eliminate the need for manual tuning and avoid forced feature alignment. Comprehensive experiments show that OGP-Net outperforms state-of-the-art models in multi-modal and missing modality scenarios across three public benchmarking datasets. It achieves quicker convergence and learns efficiently from limited training samples.

Physics-Informed Neural Networks (PINNs) offer a promising approach to simulating physical systems. Still, their application is limited by optimization challenges, mainly due to the lack of activation functions that generalize well across several physical systems. Existing activation functions often lack such flexibility and generalization power. To address this issue, we introduce Rational Exponential Activation (REAct), a generalized form of tanh consisting of four learnable shape parameters. Experiments show that REAct outperforms many standard and benchmark activations, achieving an MSE three orders of magnitude lower than tanh on heat problems and generalizing well to finer grids and points beyond the training domain. It also excels at function approximation tasks and improves noise rejection in inverse problems, leading to more accurate parameter estimates across varying noise levels.

Underwater mine-like object (MLO) detection with deep learning suffers from limitations due to the scarcity of real-world side-scan sonar (SSS) data. This scarcity leads to overfitting, where models perform well on training data but poorly on unseen data. In this letter, we propose a synthetic to real (Syn2Real) domain generalization approach using diffusion models to address this challenge. Synthetic data generated by DDPM and DDIM models effectively enhances the training dataset. The residual noise in the final sampled images improves the model’s ability to generalize to real-world data with inherent noise and high variation. The baseline mask-region-based convolutional neural network (RCNN) model when trained on a combination of synthetic and original SSS training datasets, exhibited approximately a 35% increase in average precision (AP) compared to being trained solely on the original training data. This significant improvement highlights the potential of Syn2Real domain generalization for underwater mine detection.

This paper proposes a Conflict-aware Resource-Efficient Decentralized Sequential planner (CREDS) for early wildfire mitigation using multiple heterogeneous Unmanned Aerial Vehicles (UAVs). Multi-UAV wildfire management scenarios are non-stationary, with spatially clustered dynamically spreading fires, potential pop-up fires, and partial observability due to limited UAV numbers and sensing range. The objective of CREDS is to detect and sequentially mitigate all growing fires as Single-UAV Tasks (SUT) while adhering to the physical constraints of UAV. CREDS minimizes biodiversity loss through rapid UAV intervention and promotes efficient resource utilization by avoiding complex multi-UAV coordination. CREDS employs a three-phased approach, beginning with fire detection using a search algorithm, followed by local trajectory generation using the auction-based Resource-Efficient Decentralized Sequential planner (REDS), incorporating the novel non-stationary cost function, the Deadline-Prioritized Mitigation Cost (DPMC). Finally, a conflict-aware consensus algorithm resolves conflicts to determine a global trajectory for spatiotemporal mitigation. The performance evaluation of the CREDS for partial and full observability conditions with both heterogeneous and homogeneous UAV teams for different fires-to-UAV ratios demonstrates a 100% success rate for ratios up to 4 and a high success rate for the critical ratio of 5, outperforming baselines. Heterogeneous UAV teams outperform homogeneous teams in handling heterogeneous deadlines of SUT mitigation. CREDS exhibits scalability and 100% convergence, demonstrating robustness against potential deadlock assignments, enhancing its success rate compared to the baseline approaches.Note to Practitioners—Practical wildfire scenarios often involve unknown clusters of rapidly evolving fires that exceed the available firefighting resources. Wildfire scenarios often involve vast areas and limited sensor capabilities of UAVs, resulting in partial information about the environment. When the number of fires exceeds the number of UAVs, decentralized sequential action becomes necessary. Early wildfire mitigation, focusing on containing fires within the quenching capability of a single UAV, is crucial for minimizing damage and efficient resource utilization by avoiding complex multi-UAV coordination. The approach of single UAV mitigation introduces physical constraints and deadlines for initiating mitigation efforts. The deadlines vary based on factors like fire area, spread rate, and quench rate. The computation of a quenching sequence with efficient prioritization of deadlines ensures mission success and reduction in the total destroyed area. The challenges involved in wildfire scenarios necessitate a three-phased framework: a search stage to locate fires using thermal sensors and cameras, a Resource-Efficient Decentralized Sequential planner to assign local trajectories (mitigation sequence) for each UAV, and a conflict resolution stage to ensure smooth operation by resolving potential conflicts between UAV trajectories. CREDS prioritizes deadlines and achieves successful missions even when fires outnumber UAVs by five times. Additionally, heterogeneous UAV teams with diverse quench and speed capabilities outperform homogeneous teams and traditional methods focused on execution time. This approach is well-suited for scenarios with spatially distributed, dynamic targets with diverse deadlines.

The safety and efficiency of an Autonomous Vehicle (AV) manoeuvre planning heavily depend on the future trajectories of surrounding vehicles. If an AV can predict its surrounding vehicles’ future trajectories, it can make safe and efficient manoeuvre decisions. In this paper, we present a Deep Spatio-Temporal Context-Aware decision Network (DST-CAN) for predictive manoeuvre decisions for AVs on highways. DST-CAN has two main components, namely spatio-temporal context-aware map generator and predictive manoeuvre decisions engine. DST-CAN employ a memory neuron network to predict the future trajectories of its surrounding vehicles. Using look-ahead prediction and past actual trajectories, a spatio-temporal context-aware probability occupancy map is generated. These context-aware maps as input to a decision engine generate a safe and efficient manoeuvre decision. Here, CNN helps extract feature space, and two fully connected network generates longitudinal and lateral manoeuvre decisions. Performance evaluation of DST-CAN has been carried out using two publicly available NGSIM US-101 and I-80 highway datasets. A traffic rule is defined to generate ground truths for these datasets in addition to human decisions. Two DST-CAN models are trained using imitation learning with human driving decisions from actual traffic data and rule-based ground truth decisions. The performances of the DST-CAN models are compared with the state-of-the-art Convolutional Social-LSTM (CS-LSTM) models for manoeuvre prediction. The results clearly indicate that the context-aware maps help DST-CAN to predict the decision accurately over CS-LSTM. Further, an ablation study has been carried out to understand the effect of prediction horizons of performance and a robustness study to understand the near collision scenarios over actual traffic observations. The context-aware map with a 3 second prediction horizon is robust against near collision.

In this article, a priority-based dynamic resource Allocation with decentralized Multitask assignment (P-DREAM) approach is presented to protect a territory from highly manoeuvring intruders. In the first part, static optimization problems are formulated to compute the following parameters of the perimeter defense problem; the number of reserve stations, their locations, the priority region, the monitoring region, and the minimum number of defenders required for the monitoring purpose. The concept of a prioritized intruder is proposed here to identify and handle those critical intruders (computed based on the velocity ratio and location) to be tackled on a priority basis. The computed priority region helps to assign reserve defenders sufficiently earlier such that they can neutralize the prioritized intruders. The monitoring region defines the minimum region to be monitored and is sufficient enough to handle the intruders. In the second part, the earlier developed DREAM approach is modified to incorporate the priority of an intruder. The proposed P-DREAM approach assigns the defenders to the prioritized intruders as the first task. A convex territory protection problem is simulated to illustrate the P-DREAM approach. It involves the computation of static parameters and solving the prioritized task assignments with dynamic resource allocation. Monte-Carlo results were conducted to verify the performance of P-DREAM, and the results clearly show that the P-DREAM approach can protect the territory with consistent performance against highly manoeuvring intruders.

Recent advancements in motion planning for Autonomous Vehicles (AVs) show great promise in using expert driver behaviors in non-stationary driving environments. However, learning only through expert drivers needs more generalizability to recover from domain shifts and near-failure scenarios due to the dynamic behavior of traffic participants and weather conditions. A deep Graph-based Prediction and Planning Policy Network (GP3Net) framework is proposed for non-stationary environments that encodes the interactions between traffic participants with contextual information and provides a decision for safe maneuver for AV. A spatio-temporal graph models the interactions between traffic participants for predicting the future trajectories of those participants. The predicted trajectories are utilized to generate a future occupancy map around the AV with uncertainties embedded to anticipate the evolving non-stationary driving environments. Then the contextual information and future occupancy maps are input to the policy network of the GP3Net framework and trained using Proximal Policy Optimization (PPO) algorithm. The proposed GP3Net performance is evaluated on standard CARLA benchmarking scenarios with domain shifts of traffic patterns (urban, highway, and mixed). The results show that the GP3Net outperforms previous state-of-the-art imitation learning-based planning models for different towns. Further, in unseen new weather conditions, GP3Net completes the desired route with fewer traffic infractions. Finally, the results emphasize the advantage of including the prediction module to enhance safety measures in non-stationary environments.

This paper addresses early wildfire management using a team of UAV s for the mitigation of fires. The early detection and mitigation systems help in alleviating the destruction with reduced resource utilization. A Genetic Algorithm-based Routing and Scheduling with Time constraints (GARST) is proposed to find the shortest schedule route to mitigate the fires as Single UAV Tasks (SUT). The objective of GARST is to compute the route and schedule of the UAVs so that the UAVS reach the assigned fire locations before the fire becomes a Multi UAV Task (MUT) and completely quench the fire using the extinguisher. The fitness function used for the genetic algorithm is the total quench time for mitigation of total fires. The selection, crossover, mutation operators, and elitist strategies collectively ensure the exploration and exploitation of the solution space, maintaining genetic diversity, preventing premature convergence, and preserving high-performing individuals for the effective optimization of solutions. The GARST effectively addresses the challenges posed by the NP-complete problem of routing and scheduling for growing tasks with time constraints. The GARST is able to handle infeasible scenarios effectively, contributing to the overall optimization of the wildfire management system.

Deep neural networks have shown exemplary performance on semantic scene understanding tasks on source domains, but due to the absence of style diversity during training, enhancing performance on unseen target domains using only single source domain data remains a challenging task. Generation of simulated data is a feasible alternative to retrieving large style-diverse real-world datasets as it is a cumbersome and budget-intensive process. However, the large domain-specific inconsistencies between simulated and real-world data pose a significant generalization challenge in semantic segmentation. In this work, to alleviate this problem, we propose a novel Multi-Resolution Feature Perturbation (MRFP) technique to randomize domain-specific fine-grained features and perturb style of coarse features. Our experimental results on various urban-scene segmentation datasets clearly indicate that, along with the perturbation of style-information, perturbation of fine-feature components is paramount to learn domain invariant robust feature maps for semantic segmentation models. MRFP is a simple and computationally efficient, transferable module with no additional learnable parameters or objective functions, that helps state-of-the-art deep neural networks to learn robust domain invariant features for simulation-to-real semantic segmentation. Code is available at https://github.com/airl-iisc/MRFP.

While instance segmentation models excel at object detection in satellite imagery, their performance drops when applied to slant-angled aerial images due to occlusion and scale variation. This is mainly caused by a lack of training data for such diverse viewpoints and scales. To address this limitation, we propose the Sim2Real-based Multi-Scale Context Mask-RCNN (MSC-RCNN) network, specifically designed for slant-angled aerial imagery. Sim2Real-based transfer learning is adapted to compensate for the limited availability of real-world slant-angle training data. A synthetic dataset is generated using Unreal Engine, detailing the methodology of replicating the real-world scene, for producing diverse slant-angle drone datasets with various weather conditions and backgrounds. The model leverages two distinct feature pyramid backbones, with one incorporating dilated convolutions to address large-scale objects and the other optimized for regular convolutions. Their outputs are fused to effectively detect objects across various scales and angles. Through experiments, it was demonstrated that incorporating this synthetic data significantly reduces reliance on real data while maintaining high mean Average Precision (mAP) scores. Compared to the baseline Mask R-CNN, the proposed approach with Sim2Real adaptation and the MSC-RCNN architecture achieves a remarkable 7.6% performance improvement in instance segmentation accuracy with only a 6% increase in model size. Code can be found at: https://github.com/MSC-RCNN

Enhancing the generalization capacity for semantic segmentation of aerial perception systems for safety-critical applications is vital, especially for environments with low-light and adverse conditions. Multi-spectral fusion techniques aim to maintain the merits of electro-optical (EO) and infrared (IR) images, e.g., retaining low-level features and capturing detailed textures from both modalities. However, these techniques encounter limitations when faced with scenarios involving missing modalities, especially during inference when only IR images are available. In this paper, we propose a novel spectral-based knowledge distillation architecture known as SKD-Net to improve the performance of deep learning models for missing modality scenarios for semantic segmentation tasks. In this architecture, we make use of Gated Spectral Unit to combine information from both modalities. SKD-Net aims to extract valuable semantic information from EO images while preserving spectral knowledge from the IR images within the feature space. The model retains the style information in the shallow layers while simultaneously fusing the high-level semantic context obtained from EO and IR images to improve the feature generation capacity when dealing with only IR images during inference. SKD-Net outperforms state-of-the-art multi-modal fusion and distillation models by 2.8% on average in scenarios with missing modalities when using only IR data during inference in two public benchmarking datasets. This performance increase is achieved without additional computational costs compared to the baseline segmentation models.

The scarcity of annotated infrared (IR) image datasets limits deep learning networks from achieving performances comparable to those achieved with RGB data. To address this, we introduce a novel semi-supervised RGB-to-IR Image-to-Image Translation model (SSL-RGB2IR) that generates synthetic IR data from RGB images. Our model effectively preserves the IR characteristics in the generated images from both synthetic and real-world data. Compared to existing image-to-image translation techniques, training models on this generated IR data significantly improves performance in downstream tasks like segmentation and detection. Notably, in sim-to-real transfer, the segmentation model trained on SSL-RGB2IR generated IR images outperforms baselines and other Image-to-Image (I2I) models. Furthermore, for real-world applications utilizing EO/IR fusion images, this approach solves the well-known challenge of co-registering EO and IR images, which often have inherent misalignment’s due to differing sensor characteristics. Our code is available at https://github.com/prahlad-anand/ssl-rgb2ir https://github.com/prahlad-anand/ssl-rgb2ir.

Autonomous Vehicle (AV) decision-making in ur-ban environments is inherently challenging due to the dynamic interactions with surrounding vehicles. For safe planning, AV/ego must understand the weightage of various spatiotemporal interactions in a scene. Contemporary works use colos-sal transformer architectures to encode interactions mainly for trajectory prediction, resulting in increased computational complexity. To address this issue without compromising spatiotemporal understanding and performance, we propose the simple Deep Attention Driven Reinforcement Learning (DAD-RL) framework, which dynamically assigns and incorporates the significance of surrounding vehicles into the ego's RL-driven decision-making process. We introduce an AV-centric spatiotemporal attention encoding (STAE) mechanism for learning the dynamic interactions with different surrounding vehicles. To understand map and route context, we employ a context encoder to extract features from context maps. The spatiotemporal representations combined with contextual encoding provide a comprehensive state representation. The resulting model is trained using the Soft-Actor Critic (SAC) algorithm. We evaluate the proposed framework on the SMARTS urban benchmarking scenarios without traffic signals to demonstrate that DAD-RL outperforms recent state-of-the-art methods. Furthermore, an ablation study underscores the importance of the context-encoder and spatiotemporal attention encoder in achieving superior performance.

Energy Internet exploits a network of sparsely in-terconnected prosumers and maximizes the peer-to-peer power exchanges by routing power through intermediate peers. A novel graph theory-based deterministic approach for P2P power routing that aids efficient distance-based prosumer matching while addressing limited connectivity challenges is presented here. The proposed Optimal-Walk Multi-path Power Routing (OMPR) approach consists of two steps. First, an optimal-walk connectivity algorithm calculates the shortest path between all the nodes. Leveraging this data, in the second part, the Multi-path Power Routing (MPR) algorithm, identifies optimal paths for handling multiple simultaneous P2P exchanges. In the MPR algorithm, the power scheduling and routing are formulated and solved as a Multi-path Power Scheduling Optimization (MPSO) problem, which is compared as both linear and nonlinear programming. Performance evaluation of the OMPR approach demonstrates the algorithm's scalability and capability to handle complex scenarios efficiently, including rerouting upon connectivity disruptions. The linear MPSO formulation within the OMPR facilitates real-time implementation compared to nonlinear MPSO by solving an 18-walk P2P exchange in a 100-node community in 0.18 seconds.

Multi-label learning and continual multi-label learning are crucial challenges in machine learning, particularly in handling complex data with multiple overlapping labels over time. Recent research works try to tackle the effect of data imbalance as it makes multi-label learning more challening. This work introduces an adaptive margin spiking neural net-work (AM-SNN) architecture coupled with a novel imbalance-sensitive loss function designed to enhance robustness against class imbalance in these settings. AM-SNN employs two output layers: one for predictions and another for margin values, with a unique loss function leveraging cosine similarity between predictions and ground truths to improve confidence in the model's predictions. Experiments show that AM-SNNs trained with the proposed loss function outperform state-of-the-art loss functions on metrics such as the imbalance-weighted F1 score and the F1 score for the most imbalanced class on several multi-label learning datasets. In continual multi-label learning, AM-SNNs surpass Bipolar SNNs and the CIFDM benchmark on large datasets - Birds, Human, and Eukaryote.

Autonomous Underwater Vehicles (AUVs) play a key role in modern marine exploration. The effective deployment of AUVs relies on accurately determining their dynamics, which can be modeled using system identification approaches. Data-driven approaches for system identification of AUV sub-systems is an open research area. In this paper, we present a data-driven approach for system identification of the depth subsystem of AUVs using Physics Informed Neural Networks (PINNs) and Sparse Identification of Non-Linear Dynamics (SINDy). The AUV depth subsystem is excited with two different input profiles and the resulting states are used for system identification. Both SINDy and PINN show high accuracy in the estimated coefficient values with noisy measurements. The root mean square error (RMSE) and the mean absolute error (MAE) quantify the closeness between the actual states and those obtained from the system using the estimated coefficient values. In addition, the percentage error in the identified coefficient values is reported. The results highlight that PINNs obtain the most remarkable performance with multi-step inputs, whereas SINDy excels with sine inputs. For multi-step input, PINN yields MSE and RMSE values of the order 10−2for states, whereas SINDy achieves values around of the order 10−3for sinusoidal input.

This article presents a new approach for providing an interpretation for a spiking neural network classifier by transforming it to a multiclass additive model. The spiking classifier is a multiclass synaptic efficacy function-based leaky-integrate-fire neuron (Mc-SEFRON) classifier. As a first step, the SEFRON classifier for binary classification is extended to handle multiclass classification problems. Next, a new method is presented to transform the temporally distributed weights in a fully trained Mc-SEFRON classifier to shape functions in the feature space. A composite of these shape functions results in an interpretable classifier, namely, a directly interpretable multiclass additive model (DIMA). The interpretations of DIMA are also demonstrated using the multiclass Iris dataset. Further, the performances of both the Mc-SEFRON and DIMA classifiers are evaluated on ten benchmark datasets from the UCI machine learning repository and compared with the other state-of-the-art spiking neural classifiers. The performance study results show that Mc-SEFRON produces similar or better performances than other spiking neural classifiers with an added benefit of interpretability through DIMA. Furthermore, the minor differences in accuracies between Mc-SEFRON and DIMA indicate the reliability of the DIMA classifier. Finally, the Mc-SEFRON and DIMA are tested on three real-world credit scoring problems, and their performances are compared with state-of-the-art results using machine learning methods. The results clearly indicate that DIMA improves the classification accuracy by up to 12% over other interpretable classifiers indicating a better quality of interpretations on the highly imbalanced credit scoring datasets.

This article proposes the multiswarm cooperative information-driven search and divide and conquer mitigation control (MSCIDC) approach for faster detection and mitigation of forest fires by reducing the loss of biodiversity, nutrients, soil moisture, and other intangible benefits. A swarm is a cooperative group of unmanned aerial vehicles (UAVs) flying together to search and quench the fire areas effectively. The multiswarm cooperative information-driven search uses a two-stage search comprising cooperative information-driven exploration and exploitation for quick/accurate detection of fire locations. The search level is selected based on the thermal sensor information about the potential fire area. The dynamic nature of swarms acquired from global regulative repulsion and merging between swarms reduces the detection and mitigation time compared to the existing methods. The local attraction among the swarm members helps the nondetector members reach the fire location faster, and divide-and-conquer mitigation control ensures a nonoverlapping fire sector allocation for all members quenching the fire. The performance of the MSCIDC has been compared with different multi-UAV methods using a simulated pine forest environment. The Monte-Carlo simulation results indicate that the MSCIDC reduces the average forest area burnt by$65\%$and mission time by$60\%$compared to the best case of the multi-UAV approaches, guaranteeing a faster and more successful mission.

This article presents a metacognitive decision-making (MDM) framework inspired by human-like metacognitive principles. The MDM framework is incorporated in unmanned aerial vehicles (UAVs) deployed for decentralized stochastic search without communication for detecting and confirming stationary targets (fixed/sudden pop-up) and dynamic targets. The UAVs are equipped with multiple sensors (varying sensing capability) and search for targets in a largely unknown area. The MDM framework consists of a metacognitive component and a self-cognitive component. The metacognitive component helps to self-regulate the search with multiple sensors addressing the issues of “which-sensor-to-use”, “when-to-switch-sensor”, and “how-to-search.” Based on the information gathered by sensors carried by each UAV, the self-cognitive component regulates different levels of stochastic search and switching levels for effective searching, where the lower levels of search aim to localize a target (detection) and the highest level of a search exploit a target (confirmation). The performance of the MDM framework with two sensors having a low accuracy for detection and increased accuracy to confirm targets is evaluated through Monte Carlo simulations and compared with six decentralized multi-UAV search algorithms (three self-cognitive searches and three self and social-cognitive-based searches). The results indicate that the MDM framework can efficiently detect and confirm targets in an unknown environment.

This paper considers the problem of cooperative localization of multiple robots under uncertainty, communicating over a partially connected, dynamic communication network and assisted by an agile landmark. Each robot owns an IMU and a relative pose sensing suite, which can get faulty due to system or environmental uncertainty, and therefore exhibit large bias in their estimation output. For the robots to localize accurately under sensor failure and system or environmental uncertainty, a novel Distributed Learning based Decentralized Cooperative Localization (DL-DCL) algorithm is proposed that involves real-time learning of an information fusion strategy by each robot for combining pose estimates from its own sensors as well as from those of its neighboring robots, and utilizing the moving landmark's pose information as a feedback to the learning process. Convergence analysis shows that the learning process converges exponentially under certain reasonable assumptions. Simulations involving sensor failures inducing around 40-60 times increase in the nominal bias show DL-DCL's estimation performance to be approximately 40% better than the well-known covariance-based estimate fusion methods. For the evaluation of DL-DCL's implementability and fault-tolerance capability in practice, a high-fidelity simulation is carried out in Gazebo with ROS2.

This paper addresses the problem of position estimation in UAVs operating in a cluttered environment where GPS information is unavailable. A learning-based approach is proposed that takes in the rotor RPMs and past state as input and predicts the one-step-ahead position of the UAV using a novel spectral-normalized memory neural network (SN-MNN). The spectral normalization guarantees stable and reliable prediction performance. The predicted position is transformed to the global coordinate frame (GPS), which is then fused along with the odometry of other peripheral sensors like IMU, barometer, compass, etc., using the onboard extended Kalman filter (EKF) to estimate the states of the UAV. The experimental flight data collected from an RTK-GPS facility using a micro-UAV is used to train the SN-MNN. The PX4-ECL library is used to fuse the predicted data using the SN-MNN, and the estimated position is compared with actual ground truth data. The proposed algorithm doesn't require any additional onboard sensors and is computationally light. The performance of the proposed approach is compared with the current state-of-art GPS-denied algorithms, and it can be seen that the proposed algorithm has the least RMSE for position estimates.

This paper tackles the out-of-distribution (OOD) detection problem for natural language classifiers. While the previous OOD detection methods require large-scale in-distribution (ID) training data, we attack this problem from the few-shot perspective in an unsupervised manner where the training relies on only a few samples from ID data. First, we develop various baselines for Few-shot OOD (FSOOD) detection in text classification based on the three well-known few-shot learning approaches (well-explored in the vision domain), i.e., meta-learning, metric learning, and data augmentation (DA). Then, we introduce the concept of demonstration-based data augmentation with meta and metric-learning approaches to reap the combined benefit of both approaches. A pre-trained transformer is fine-tuned on a few available ID samples in all developed methods. In tandem with this fine-tuning, an OOD detector is fitted over the ID training samples to reject the data from the unknown classes using two kinds of distance metrics, namely Mahalanobis distance and Cosine similarity. At last, we present an extensive evaluation of three ID datasets and three OOD datasets. We also perform an ablation study to analyze the impact of various components of our method.

Deep learning models operating in the complex domain are used due to their rich representation capacity. However, most of these models are either restricted to the first quadrant of the complex plane or project the complex-valued data into the real domain, causing a loss of information. This paper proposes that operating entirely in the complex domain increases the overall performance of complex-valued models. A novel, fully complex-valued learning scheme is proposed to train a Fully Complex-valued Convolutional Neural Network (FC-CNN) using a newly proposed complex-valued loss function and training strategy. Benchmarked on CIFAR-10, SVHN, and CIFAR-100, FC-CNN has a 4-10% gain compared to its real-valued counterpart, with the same number of parameters. It achieves comparable performance to state-of-the-art complex-valued models on CIFAR-10 and SVHN with fewer parameters. For the CIFAR-100 dataset, it achieves state-of-the-art performance with 25% fewer parameters. FC-CNN shows better training efficiency and much faster convergence than all the other models.

This paper develops a new approach to estimate predicted class probabilities in deep Spiking Neural Networks (SNN) that encourages faster classification. The proposed approach utilizes the temporal separation between the first spikes generated by the output neurons to estimate the predicted class probabilities which are then used with cross entropy loss for training the network. This maximizes the separation between the first spikes generated by the neuron associated with the correct class and neurons associated with other classes. Higher classification performance is obtained by maximising the tem-poral separation, which also drives the correct class neuron to spike earlier in the simulation. As a consequence, the predicted class may be determined from the first spike in the output layer, leading to quicker classification. The sensitivity factor for each neuron in the network is estimated via error-backpropagation during training. Using Spike Timing Dependent Plasticity (STDP) regulated by the estimated sensitivity factors, the network weights are updated. It results that the learning method is termed as Temporal Separation Modulated Spike Timing Dependent Plasticity (TSM-STDP). On the benchmark MNIST dataset, the performance of TSM-STDP has been assessed, and the evaluation results are compared with those of other learning methods for SNNs. Additionally, a histogram of the output layer's first spikes demonstrated that the right class neurons spiked earlier in the simulation than other class neurons, enabling faster classification. On real-world Attention Deficit Hyperactivity Disorder (ADHD) detection dataset, the effectiveness of TSM-STDP has also been assessed and compared with other available approaches. The per-formance comparison results clearly show that TSM-STDP can achieve classification performance comparable to other existing learning algorithms on benchmark and real-world datasets while requiring less time for classification.

Protein secondary structure (SS) prediction is a classic problem of computational biology and is widely used in structural characterization and to infer homology. While most SS predictors have been trained on thousands of sequences, a previous approach had developed a compact model of training proteins that used aC-Alpha, C-BetaSide Chain (CABS)-algorithm derived energy based feature representation. Here, the previous approach is extended to Deep Belief Networks (DBN). Deep learning methods are notorious for requiring large datasets and there is a wide consensus that training deep models from scratch on small datasets, works poorly. By contrast, we demonstrate a simple DBN architecture containing a single hidden layer, trained only on the CB513 dataset. Testing on an independent set of G Switch proteins improved the Q$_{3}$score of the previous compact model by almost 3%. The findings are further confirmed by comparison to several deep learning models which are trained on thousands of proteins. Finally, the DBN performance is also compared withPositionSpecificScoringMatrix (PSSM)-profile based feature representation. The importance of (i) structural information in protein feature representation and (ii) complementary small dataset learning approaches for detection of structural fold switching are demonstrated.

Designing a Convolutional Neural Networks (CNN) is a complex task and requires expert knowledge to optimize the performance and network architecture. In this paper, a novel data-driven approach is proposed to determine the architecture of CNN models. The proposed Autonomous Convolutional Neural Networks (AutoCNNThe executable code and original numerical results can be downloaded from (https://tinyurl.com/AutoCNN)) algorithm introduces data driven strategies for addition of new convolutional layers, pruning of redundant filters and training cycle optimization. AutoCNN is evaluated using MNIST, MNIST-rot-back-image, Fashion MNIST and the ADHD200 datasets to measure the performance on small datasets with varied feature distributions. The results indicate that AutoCNN optimizes the CNN network architecture and helps maximise the classification performance. The data-driven network determination approach introduced in this paper was found to not only provides competitive performance similar to existing evolutionary computation based network determination algorithms in literature, but was found to be an effective optimization tool to improve the performance of existing CNN architectures. Further, the AutoCNN was found to highly immune to noise in the dataset and has proven to be effective method to transfer knowledge between related datasets. Therefore, the AutoCNN is a highly versatile CNN architecture determination tool that has a wide range of applications in the field of autonomous driving, medical image analysis, image enhancement, camera based security monitoring and image based fault detection

In this article, we propose a bilevel segmentation framework with metacognitive learning (BS-McL) to detect power lines with an RGB camera mounted on an unmanned aerial vehicle (UAV) platform. The proposed framework consists of two levels based on spectral and spatial techniques. In the first level, spectral classification is carried out using the McL method, which is an evolving online learning neural network architecture. Due to similarities in spectral intensities, few nonpower line pixels are grouped along with power line pixels. The nonpower line pixels are removed by spatial segmentation in the second level. The second level includes morphological operations such as geometric features (shape and density indices), which are applied to detect the power lines. The processing steps of BS-McL are illustrated using a synthetic image of size$9 \times 6$pixels. Also, two datasets consisting of 64 images with varying backgrounds, different locations, and dimensions of power lines are used to demonstrate the performance of the proposed BS-McL. The obtained results for BS-McL are compared with five commonly used methods. For both datasets, the efficiency of the BS-McL for power line extraction is better than for the methods used for comparison. Furthermore, the trained knowledge from our experimental set-up (Dataset 1: suburban scene) can be transferred to another dataset that is available publicly (Dataset 2: urban and mountain scenes) if the power line spectral values are in relevance with the distribution in the training dataset. The proposed approach BS-McL is based on online learning with a self-adaptive architecture, which provides improved generalization ability.

This article provides a new decentralized approach to solve a perimeter defense problem (PDP). In a typical PDP, many intruders try to enter a territory, and a group of defenders operating both inside and on the perimeter protects the territory by capturing the intruders on the perimeter. The objective of the defenders is to detect and capture the intruders before they enter the territory. Defenders sense the intruders independently and compute their trajectories to capture all the intruders in a cooperative way. Each intruder is estimated to reach a specific location on the perimeter at a specific time, and this is considered as a spatiotemporal task to be handled by a defender. At any given time, the PDP is converted to a decentralized multirobot spatiotemporal multitask assignment (DMRST-MTA) problem. The cost of executing a task for a defender is defined by a composite cost function that includes both the spatial and temporal cost components. In this article, a modified decentralized consensus-based bundle algorithm is presented to solve the above spatiotemporal multitask assignment problem. The performance evaluation of the proposed approach is presented based on Monte Carlo studies, and the results show the effectiveness of the proposed approach under different scenarios. The robustness studies also show that the proposed approach is robust against uncertainties in the heading angles of the intruders. The performance comparison with the decentralized adaptive partitioning approach for the various arrival distributions of intruders clearly shows that DMRST-MTA is efficient.

The plants of nano air vehicles (NAVs) are generally unstable, adversely coupled, and uncertain. Besides, the autopilot hardware of a NAV has limited sensing and computational capabilities. Hence, these vehicles need a single controller referred to as robust simultaneously stabilizing decoupling (RSSD) output feedback controller that achieves simultaneous stabilization (SS), desired decoupling, robustness, and performance for a finite set of unstable multi-input–multioutput adversely coupled uncertain plants. To synthesize an RSSD output feedback controller, a new method that is based on a central plant is proposed in this article. Given a finite set of plants for SS, we considered a plant in this set that has the smallest maximum$v-$gap metric as the central plant. Following this, the sufficient condition for the existence of a simultaneous stabilizing controller associated with such a plant is described. The decoupling feature is then appended to this controller using the properties of the eigenstructure assignment method. Afterward, the sufficient conditions for the existence of an RSSD output feedback controller are obtained. Using these sufficient conditions, a new optimization problem for the synthesis of an RSSD output feedback controller is formulated. To solve this optimization problem, a new genetic algorithm-based offline iterative algorithm is developed. The effectiveness of this iterative algorithm is then demonstrated by generating an RSSD controller for a fixed-wing NAV. The performance of this controller is validated through numerical and hardware-in-the-loop simulations.

Prognostication of vehicle trajectories in unknown environments is intrinsically a challenging and difficult problem to solve. The behavior of such vehicles is highly influenced by surrounding traffic, road conditions, and rogue participants present in the environment. Moreover, the presence of pedestrians, traffic lights, stop signs, etc., makes it much harder to infer the behavior of various traffic agents. This paper attempts to solve the problem of spatio-temporal look-ahead trajectory prediction using a novel recurrent neural network called the Memory Neuron Network. The Memory Neuron Network (MNN) attempts to capture the input-output relationship between the past positions and the future positions of the traffic agents. The proposed prediction model is computationally less intensive and has a simple architecture as compared to other deep learning models that utilize LSTMs and GRUs. It is then evaluated on the publicly available NGSIM dataset and its performance is compared with several state-of-art algorithms. Additionally, the performance is also evaluated on a custom synthetic dataset generated from the CARLA simulator. It is seen that the proposed model outperforms the existing state-of-art algorithms. Finally, the model is integrated with the CARLA simulator to test its robustness in real-time traffic scenarios.

When it comes to time-series forecasting, it is crucial to learn the intricate temporal relationship between past and future, and historical information is of paramount importance for this purpose. Traditional neuro-fuzzy systems generally resort to an empirical (offline) method to determine the number of past instances (i.e., historical information) required for a particular model, hence often unsuitable in an online time-series scenario. In this article, we propose a Bayesian neuro-fuzzy inference system (BaNFIS), where the temporal dependence on past instances is estimated with an online Bayesian probabilistic mechanism, and the uncertainty associated with real-world data is handled by the fuzzy inference system. The BaNFIS retains historical information only as per necessity and employs it in two ways: globally or locally. Moreover, an online learning method is employed here to update the BaNFIS parameters. Hence, the BaNFIS is able to capture both the system dynamics and uncertainty efficiently in an online manner. Three real-world time-series problems are employed here to evaluate the online performance of the BaNFIS compared to seven state-of-the-art neuro-fuzzy methods both under standard train–test and prequential test–train protocols. Numerical results clearly indicate that the BaNFIS provides a statistically improved prediction performance than its peers in terms of accuracy.

In the past few years, kernel-based one-class extreme learning machine (ELM) receives quite a lot of attention by researchers for offline/batch learning due to its noniterative and fast learning capability. This paper extends this concept for adaptive online learning with regularized kernel-based one-class ELM classifiers for detection of outliers, and are collectively referred to as ORK-OCELM. Two frameworks, viz., boundary and reconstruction, are presented to detect the target class in ORK-OCELM. The kernel hyperplane-based baseline one-class ELM model considers whole data in a single chunk, however, the proposed one-class classifiers are adapted in an online fashion from the stream of training samples. The performance of ORK-OCELM is evaluated on a standard benchmark as well as synthetic datasets for both types of environments, i.e., stationary and nonstationary. While evaluating on stationary datasets, these classifiers are compared against batch learning-based one-class classifiers. Similarly, while evaluating on nonstationary datasets, the comparison is done with incremental learning-based online one-class classifiers. The results indicate that the proposed classifiers yield better or similar outcomes for both. In the nonstationary dataset evaluation, adaptability of the proposed classifiers in a changing environment is also demonstrated. It is further shown that the proposed classifiers have large stream data handling capability even under limited system memory. Moreover, the proposed classifiers gain significant time improvement compared to traditional online one-class classifiers (in all aspects of training and testing). A faster learning ability of the proposed classifiers makes them more suitable for real-time anomaly detection.

This paper proposes a Distributed Coding Spiking Neural Network (DC-SNN) with a self-regulated learning algorithm to deal with pattern classification problems. DC-SNN employs two hidden layers. First hidden layer has receptive field neurons that convert the real-valued input features to spike patterns and the second hidden layer employs LIF neurons with inhibitory interconnections. The second hidden layer has been termed as the distributed coding layer in the rest of the paper. The inhibitory interconnections in distributed coding layer will ensure that each neuron in this layer learns a distinct spike pattern from input feature space. The synaptic weights between layers and the weights of lateral inhibitory connections are learned using a self-regulated learning algorithm. Self-regulation identifies neurons for updating in the output layer and distributed coding layer and also adapts the learning rate based on the temporal separation between spikes in the output layer. It also skips learning from samples which are correctly classified with higher temporal separation and hence prevents over-training. The detailed performance comparisons of DC-SNN with other algorithms for SNNs in the literature using six benchmark data set from the UCI machine learning repository has been presented. Further, the performance of DC-SNN is evaluated on a real-world brain computer interface problem for classification of electroencephalogram (EEG) signals recorded during motor-imagery tasks. The results clearly indicate that the proposed DC-SNN architecture provides slightly better generalization ability and is suitable for deep spiking networks.

When it comes to time series forecasting, lag in the predicted sequence can be a predominant issue. Unfortunately, this is often overlooked in most of the time series literature as this does not contribute to a high prediction error (i.e. MSE). However, it leads to a rather poor forecast in terms of movement prediction in time series. In this article, we tackle this basic problem with a novel trend driven mechanism. Trend, defined as the inherent pattern of the data, is extracted here and utilized next to perform a lag-free forecasting. We propose a generic and light Dual Network Solution (DNS), where the first network predicts the trend and the second network utilizes that predicted trend along with its historical information to capture the dynamical behavior of the time series efficiently. DNS exhibits a substantially improved (≈10% better) performance compared to more complex and resource-intensive state-of-the- art algorithms in large scale regression problems. Apart from the traditional Mean Squared Error (MSE), we also propose a new Movement Prediction Metric or MPM (for detection of lag in time series) as a new complementary performance metric to evaluate the efficacy of DNS better.

In this paper, we propose a mission aware motion planning (MAP) framework for a swarm of autonomous unmanned ground vehicles (UGVs) or mobile stations in an uncertain environment for efficient supply of resources/services to unmanned aerial vehicles (UAVs) performing a specific mission. The MAP framework consists of two levels, namely, centralized mission planning and decentralized motion planning. On the first level, the centralized mission planning algorithm estimates the density of UAV in a given environment for determining the number of UGVs and their initial operating location. In the subsequent level, a decentralized motion planning algorithm which provides a closed-form expression for velocity command using adaptive density estimation has been proposed. Further, the physical and geographical constraints are integrated into motion planning. A Monte-Carlo simulation is performed to evaluate the advantages of the MAP over distributed stationary stations (DSSs) often used in the literature. The obtained results clearly indicate that in comparison with DSS, MAP reduces the average distance traveled by UAVs about 20%, reduces the loss of mission time by 90 s per interruption and power loss by 3 dB.

This article presents a novel solution technique for scheduling multi-energy system (MES) in a commercial urban building to perform price-based demand response and reduce energy costs. The MES scheduling problem is formulated as a mixed integer nonlinear program (MINLP), a nonconvex NP-hard problem with uncertainties due to renewable generation and demand. A model predictive control approach is used to handle the uncertainties and price variations. This in-turn requires solving a time-coupled multitime step MINLP during each time-epoch, which is computationally intensive. This investigation proposes an approach called the scenario-based branch-and-bound (SB3), a light-weight solver to reduce the computational complexity. It combines the simplicity of convex programs with the ability of meta-heuristic techniques to handle complex nonlinear problems. The performance of the SB3 solver is validated in the Cleantech building, Singapore and the results demonstrate that the proposed algorithm reduces energy cost by about 17.26% and 22.46% as against solving a multi-time step heuristic optimization model.

Analysis of resting state - functional Magnetic Resonance Imaging (rs-fMRI) data has been a challenging problem due to a high homogeneity, large intra-class variability, limited samples and difference in acquisition technologies/techniques. These issues are predominant in the case of Attention Deficit Hyperactivity Disorder (ADHD). In this paper, we propose a new Deep Transformation Method (DTM) that extracts the discriminant latent feature space from rsfMRI and projects it in the subsequent layer for classification of rs-fMRI data. The hidden transformation layer in DTM projects the original rs-fMRI data into a new space using the learning policy and extracts the spatio-temporal correlations of the functional activities as a latent feature space. The subsequent convolution and decision layers transform the latent feature space into high-level features and provide accurate classification. The performance of DTM has been evaluated using the ADHD200 rs-fMRI benchmark data with crossvalidation. The results show that the proposed DTM achieves a mean classification accuracy of 70.36% and an improvement of 8.25% on the state of the art methodologies was observed. The improvement is due to concurrent analysis of the spatio-temporal correlations between the different regions of the brain and can be easily extended to study other cognitive disorders using rs-fMRI. Further, brain network analysis has been studied to identify the difference in functional activities and the corresponding regions behind cognitive symptoms in ADHD.

Two of the major challenges associated with time series modelling are handling uncertainty present in the data and tracing its dynamical behaviour. A Recurrent Interval Type 2 Fuzzy Inference System or RIT2FIS is presented in this paper. RIT2FIS adopts an interval type 2 fuzzy inference mechanism for superior handling of uncertainty. The memory neurons employed in its hidden and output layer, retain the temporal information, making RIT2FIS highly proficient in tracing system dynamics at a granular level. RIT2FIS also benefits from incorporating a k-means algorithm inspired approach to cluster the data in an unsupervised manner. An 'Elbow Method' is utilized next to determine the optimal clustering which is then employed as the optimal fuzzy rule base for RIT2FIS, eliminating the necessity of expert knowledge for fuzzy initiation. The antecedent and consequent parameters of RIT2FIS are updated using a gradient descent based back-propagation through time algorithm where the learning is made self-regulatory to avoid over-fitting and ensure generalization. Performance of RIT2FIS is evaluated against popular neuro-fuzzy methods on different benchmark and real-world time series problems which distinctly indicates an improved accuracy and a parsimonious rule base.

This paper presents an Oxyrrhis Marina-inspired search and dynamic formation control (OMS-DFC) framework for multi-unmanned aerial vehicle (UAV) systems to efficiently search and neutralize a dynamic target (forest fire) in an unknown/uncertain environment. The OMS-DFC framework consists of two stages, viz., the target identification stage without communication between UAVs and the mitigation stage with restricted communication. In the first stage, each UAV adapts proposed OMS with three levels to select between Levy flight, Brownian search, and directionally driven Brownian (DDB) search for accurate target identification (“fire location”). The selection of each level is based on the available sensor information about the possible fire location. In the second stage, the UAVs that identified a fire location fly in a dynamic formation to quench the fire using water. The proposed formation is achieved through decentralized control, where a UAV computes the control action based on the fire profile and also the angular position and angular separation with its succeeding neighbor. The proposed formation control law guarantees asymptotic convergence to the desired time-varying angular position profile of UAVs based on the nature of fire spread (circular/elliptical). To evaluate the performance of the proposed OMS-DFC for the multi-UAV system, a search and fire quenching mission in a typical pine forest is simulated. A Monte Carlo simulation study is conducted to evaluate the average performance of the proposed OMS-DFC-based multi-UAV mission, and the results clearly highlight the advantages of the proposed OMS-DFC in forest firefighting.

This paper presents a new learning algorithm developed for a three layered spiking neural network for pattern classification problems. The learning algorithm maximizes the interclass margin and is referred to as the two stage margin maximization spiking neural network (TMM-SNN). In the structure learning stage, the learning algorithm completely evolves the hidden layer neurons in the first epoch. Further, TMM-SNN updates the weights of the hidden neurons for multiple epochs using the newly developed normalized membrane potential learning rule such that the interclass margins (based on the response of hidden neurons) are maximized. The normalized membrane potential learning rule considers both the local information in the spike train generated by a presynaptic neuron and the existing knowledge (synaptic weights) stored in the network to update the synaptic weights. After the first stage, the number of hidden neurons and their parameters are not updated. In the output weights learning stage, TMM-SNN updates the weights of the output layer neurons for multiple epochs to maximize the interclass margins (based on the response of output neurons). Performance of TMM-SNN is evaluated using ten benchmark data sets from the UCI machine learning repository. Statistical performance comparison of TMM-SNN with other existing learning algorithms for SNNs is conducted using the nonparametric Friedman test followed by a pairwise comparison using the Fisher's least significant difference method. The results clearly indicate that TMM-SNN achieves better generalization performance in comparison to other algorithms.

In this paper, a cognitive decision-making architecture for dynamic airspace sectorization (CDAS) to handle increasing traffic flow and provide an efficient decision making process for operations is presented. The main objective of CDAS is to determine optimal 3-D sector shapes such that the conflicting workloads of air traffic controllers are balanced along with marginal changes to sector shapes over a time horizon. CDAS broadly comprises two major components, namely, a cognitive engine and a metacognitive decision maker. The cognitive engine includes an airspace sectorization model and a multi-objective optimization solver. The problem of 3-D dynamic airspace resectorization is cast as a multi-objective optimization problem with safety constraints and is solved using the non-dominated sorting genetic algorithm II. The metacognitive decision maker utilizes the Pareto-optimal solutions obtained from the cognitive engine along with the air traffic control requirements and predicted traffic pattern to identify the best solution that can be implemented along with a rule to decide on when-to-do a resectorization when needed. A detailed performance evaluation of CDAS is presented using the actual flight data over the Singapore flight information region. The results clearly indicate that CDAS provides an efficient dynamic sectorization solution over the existing solution of split-and-merge of a specific sector to handle heavy traffic.

This paper presents a new time-varying long-term Synaptic Efficacy Function-based leaky-integrate-and-fire neuRON model, referred to as SEFRON and its supervised learning rule for pattern classification problems. The time-varying synaptic efficacy function is represented by a sum of amplitude modulated Gaussian distribution functions located at different times. For a given pattern, the SEFRON's learning rule determines the changes in the amplitudes of weights at selected presynaptic spike times by minimizing a new error function reflecting the differences between the desired and actual postsynaptic firing times. Similar to the gamma-aminobutyric acid-switch phenomenon observed in a biological neuron that switches between excitatory and inhibitory postsynaptic potentials based on the physiological needs, the time-varying synapse model proposed in this paper allows the synaptic efficacy (weight) to switch signs in a continuous manner. The computational power and the functioning of SEFRON are first illustrated using a binary pattern classification problem. The detailed performance comparisons of a single SEFRON classifier with other spiking neural networks (SNNs) are also presented using four benchmark data sets from the UCI machine learning repository. The results clearly indicate that a single SEFRON provides a similar generalization performance compared to other SNNs with multiple layers and multiple neurons.

The problem of estimating the position, velocity, and acceleration of an agile target from imperfect position measurements is addressed in this paper. A discrete sliding mode observer (DSMO) is used that can handle measurement inaccuracies apart from unmeasured disturbance inputs. The target acceleration input acts as an unmeasured disturbance input to the observer. The parameters of DSMO are derived considering the worst case measurement error and unmeasured disturbance inputs. The selected parameters guarantee the convergence of the error dynamics into the boundary layer and finite state estimation error within the boundary layer. A numerical simulation study is presented for the state estimation of a sinusoidally maneuvering target. Experimental results are presented for the state estimation of a target unmanned ground vehicle (UGV), with position information obtained from a camera mounted on an unmanned air vehicle (UAV).

The problem of 3-Dimensional en-route airspace sectorization has been modelled as a multi-objective optimization by taking into account of conflicting air traffic controller workloads. In practice, air traffic controllers impose limits on these objectives, which may not be captured completely in Pareto front obtained using the multi-objective model. Hence, in this paper, we propose a preference-based multi-objective optimization model for 3-dimensional en-route sectorization. Through the use of reference point(s), the proposed model is able to find multiple solutions that satisfy the air traffic controller preference(s) faster. Population-based NSGA-II has been used to solve the preference-based 3-dimensional en-route sectorization problem. The performance of preference based sectorization is evaluated using actual flight data from the Singapore regional airspace. The results are compared with conventional multi-objective optimization model which integrates the preference as a constraint. Results indicate that the preference-based model generally performs better than the constraint-based model. Further, multiple parallel runs of preference-based optimization could provide a greater variety of choices in airspace sectorization for the air traffic controllers.

Random embedding has been applied with empirical success to large-scale black-box optimization problems with low effective dimensions. This paper proposes the EmbeddedHunter algorithm, which incorporates the technique in a hierarchical stochastic bandit setting, following the optimism in the face of uncertainty principle and breaking away from the multiple-run framework in which random embedding has been conventionally applied similar to stochastic black-box optimization solvers. Our proposition is motivated by the bounded mean variation in the objective value for a low-dimensional point projected randomly into the decision space of Lipschitz-continuous problems. In essence, the EmbeddedHunter algorithm expands optimistically a partitioning tree over a low-dimensional — equal to the effective dimension of the problem —search space based on a bounded number of random embeddings of sampled points from the low-dimensional space. In contrast to the probabilistic theoretical guarantees of multiple-run random-embedding algorithms, the finite-time analysis of the proposed algorithm presents a theoretical upper bound on the regret as a function of the algorithm's number of iterations. Furthermore, numerical experiments were conducted to validate its performance. The results show a clear performance gain over recently proposed random embedding methods for large-scale problems, provided the intrinsic dimensionality is low.

In this paper, we develop a human social intelligence inspired population-based optimization algorithm called Higher Order Cognitive Optimization (HOCO) algorithm. Each of the individuals in this HOCO possess human-like characteristics such as decision making ability, self/social-awareness, self/social belief, shared information processing, and self-regulation. These characteristics are modeled as a hierarchical inter-related structure with each layer realizing different levels of granularity. In this paper, HOCO is implemented as a three layered inter-related architecture for single-objective optimization. The main aspects of the proposed optimization technique are: (1) development of a socially intelligent optimization algorithm; (2) each individual employs their meta-cognitive as well as social meta-cognitive abilities, in addition to the cognitive abilities to attain the global optimal solution; and (3) the meta-cognitive and social meta-cognitive components self-regulate the cognitive component by adapting its strategies, such that a globally optimal solution formulation is achieved. Performance has been analyzed on six standard benchmark problems and compared with other meta-heuristic algorithms. Further, the performance on computationally expensive CEC2015 benchmark problems has also been studied. The comparison with other population based meta-heuristic approaches indicates the significance of the HOCO algorithm.

In this paper, we conduct a comprehensive study to identify the most discriminative features that address the interpersonal variability to perform efficient human emotion recognition task. We consider three commonly used feature extraction techniques, namely, the Local Binary Patterns (LBP), the Scale-Invariant Feature Transform (SIFT) and the curvelet transforms to extract features from the images on the JAFFE data set. A subset of these features is then selected using the Double Input Symmetrical Relevance (DISR), the Conditional Mutual Information Maximization (CMIM) and the minimum Redundancy maximum Relevance (mRMR) methods. The original feature sets and the subsets are then used to train a PBL-McRBFN classifier. We conduct a subject independent study with 10 cross validations on the JAFFE data set. The average performance of the PBL-McRBFN classifier with the different feature sets and subsets are compared. In general, feature selection methods used along with the feature extraction techniques help to perform emotion recognition more efficiently. It is also observed that the subset of features selected using the mRMR on the features extracted from the SIFT technique (SIFT+mRMR+PBL-McRBFN) is the most discriminative feature subset. A statistical paired t-test also ascertains this observation. We also compare the performance of the SIFT+mRMR+PBL-McRBFN with the other results in the literature for this problem. Performance comparison shows that the SIFT+mRMR+PBL-McRBFN outperforms other state-of-the-art methods in the literature for this problem.

This paper addresses the problem of on-line learning for object tracking. Although a variety of techniques have been proposed in literature, a recent benchmark reveals that none of them can work well in all scenarios due to numerous practical challenges, such as illumination variations, motion blur, etc. These challenges occur at different time frames making it hard to design a tracker. In this paper, a machine learning framework for object tracking is investigated, which can integrate with a variety of feature and kernel engineering techniques in dealing with many challenges in different scenarios. By following the successful tracking-by-detection methodology, this paper proposes OMKT - On-line Multiple Kernel Tracking scheme, which attempts to tackle the object tracking task by exploring recent advances of on-line multiple kernel learning techniques in machine learning. In particular, OMKT sequentially learns the best tracker for each individual kernel via on-line Projectron++ learning, and at the same time attempts to identify the optimal combination of multiple kernel trackers using the Hedge algorithm. In contrast to many existing schemes in literature, OMKT learns both the kernel classifiers and their combination on-line, hence it can adapt to tracking changes faster. Furthermore, the projection strategy in Projectron++ alleviates the difficulty of pre-specifying a budget size for support-set. Promising experimental results on a recent benchmark reveal usefulness of OMKT.

This paper proposes a new algorithm for the multiple instance learning problem (MIL) and investigates its application for detecting Attention Deficit Hyperactive Disorder (ADHD) from resting-state functional Magnetic Resonance Imaging data. The core component of many kernel-based MIL algorithms is usually an SVM-like batch optimization framework, hence scaling to large datasets like fMRI is often difficult. On the other hand, a family of on-line kernel classification algorithms widely known as “perceptron-like” kernel classifiers demonstrate efficient and accurate solutions. This paper presents MiPAL - Multiple-instance Passive Aggressive Learning algorithm, based on such a perceptron-like kernel classifier. First, MiPAL builds a labeller by inputting negative bags into PA algorithm. Second, this labeller helps to train a separate PA classifier to predict binary class labels such that least-negative instances are regarded as positive. Due to the on-line PA algorithm's fast adaptation, the impact of invalid positive support-vectors could be attenuated by the new, accurate support-set over time. Our experimental results reveal performance gains in several MIL datasets including state-of-the-art performance in Muskl, Fox, and comparable accuracy in the preprocessed ADHD-200 dataset.

Decomposition-based evolutionary algorithms have been applied with success to multi-objective optimization problems where they are broken into several subproblems, and solutions for the original problem are recognized in a coordinated manner. Motivated by the working principle of decomposition-based methods, viz. the “divide-and-conquer” paradigm; this paper is concerned with solving black-box multi-objective problems given a finite number of function evaluations by taking inspiration from the single-objective deterministic sampling method, DIRECT. In particular, we provide a multi-objective algorithmic instance of DIRECT, which we refer to as MO-DIRECT and investigate its performance with respect to established decomposition-based multi-objective techniques. Besides its asymptotic convergence to the Pareto front and its inherent balance between exploration and exploitation of the decision space, the proposed framework is flexible enough to incorporate indicator-based techniques, albeit at the cost of a greater space complexity when compared to the evolutionary counterpart.

This paper provides an empirical analysis of the recently proposed Bayesian Multi-Scale Optimistic Optimization (BaMSOO) algorithm for solving bound-constrained black-box global optimization problems under expensive as well as cheap budgets of function evaluations. The CEC2016 and BBOB benchmarks are used in assessing the algorithm. We also compare BaMSOO with the Simultaneous Optimistic Optimization (SOO) algorithm from which BaMSOO is derived. The results indicate that BaMSOO has a comparable performance with SOO under expensive-budget settings but outperforms it with increasing evaluation budgets. Finally, we provide useful insights regarding the algorithm's relative efficiency and future directions.

This paper presents a method for a hands-off noninvasive brain computer interface (BCI) to control the movement of a quadcopter. The quadcopter model used in this paper is the AR.drone 2.0 and the non-invasive BCI device used is the Emotiv EPOC. A framework is developed to convert the raw EEG signals into commands to control the flight of quadcopter, AR.drone 2.0 through a wireless interface. The common spatial pattern algorithm is used to extract features from the EEG signals. The signals are classified using the meta-cognitive interval type-2 neuro-fuzzy inference system. The decoded intentions of the user, commands (move left, move right) are relayed to the quad over a wireless connection. Experiments show that the user is able to fly the quadcopter successfully without using hands.

In this paper, we propose an interval type-2 fuzzy inference system using Extended Kalman Filter based learning algorithm. It is referred to as IT2FIS-EKF. This algorithm realizes the Takagi-Sugeno-Kang inference mechanism in a five layered architecture. It starts with no rules and evolves the structure automatically. The sequential learning algorithm regulates the learning process by selecting appropriate learning strategies to evolve the architecture and estimate the antecedent/consequent parameters. The performance of IT2FIS-EKF is evaluated on a set of benchmark classification problems from the UCI machine learning repository. Results show superior performance of IT2FIS-EKF in comparison to other fuzzy neural networks due fully adaptive nature of learning algorithm. Further, IT2FIS-EKF is applied to a practical problem of classification of motor-imagery tasks in motor-imagery based brain computer interface (BCI). Performance is evaluated using the publicly available BCI competition dataset. Results indicate superior performance of IT2FIS-EKF making it suitable for BCI.

In this paper, we present a novel approach for Remaining Useful Life (RUL) estimation problem in prognostics using a proposed `sequential learning Meta-cognitive Regression Neural Network (McRNN) algorithm for function approximation'. The McRNN has two components, namely, a cognitive component and a meta-cognitive components. The cognitive component is an evolving single hidden layer Radial Basis Function (RBF) network with Gaussian activation functions. The meta-cognitive component present in McRNN helps to cognitive component in selecting proper samples to learn based on its current knowledge and evolve architecture automatically. The McRNN employs extended Kalman Filter (EKF) to find optimal network parameters in training. First, the performance of the proposed sequential learning McRNN algorithm has been evaluated using a set of benchmark function approximation problems and is compared with existing sequential learning algorithms. The performance results on these problems show the better performance of McRNN algorithm over the other algorithms. Next, the proposed McRNN algorithm has been applied to RUL estimation problem based on sensor data. For simulation studies, we have used Prognostics Health Management (PHM) 2008 Data Challenge data set and compared with the existing approaches based on state-of-the-art regression algorithms. The experimental results show that our proposed McRNN algorithm based approach can accurately estimate RUL of the system.

The membrane proteins are an important group of molecules whose 3-D structure is difficult to obtain experimentally. Membrane proteins are implicated as drug targets and play an important role in disease pathways. The computational structure prediction from membrane protein sequences aids understanding of the structures. The prediction of structural preferences of individual residues within a protein sequence is often used as a starting point by higher order structure prediction algorithms that predict atomic coordinates. The low number of membrane proteins relative to globular proteins is a motivation for classifiers employing a meta-cognitive framework, as they have been shown in the machine learning literature to learn from a smaller number of samples and to generalize well to new datasets. In this paper, the recently developed Projection based, Meta-cognitive Radial Basis Function Network (PBL-McRBFN) was used in the membrane protein structure prediction problem. The PBL-McRBFN consists of a cognitive component that employs a projection-based learning algorithm. The meta-cognitive component controls the architecture and learning strategies of the cognitive component. The prediction of residue preferences with respect to the membrane (inside, outside, membrane) is considered as a three-category classification problem. A dataset of transmembrane (TM) helix sequences was encoded as Position Specific Scoring Matrices (PSSM) and the performance compared with an SVM classifier. The results of the study indicate that the PBL-McRBFN classifier performs better than the SVM for the overall accuracy and is able to generalize better to the test residues than the SVM classifier.

This paper presents a BCI system which addresses the key problems of robust feature extraction, non-stationarity and subject-specific spectral filter selection. It employs the Robust Common Spatial pattern (RoCSP) feature extraction algorithm which eliminates trials affected by artifacts and discards redundant channels to improve the robustness of the CSP algorithm. Next, it handles the non-stationarity in EEG signals using the Self-Regulated Interval Type-2 Neuro-Fuzzy Inference System (SRIT2NFIS). It uses the input features generated by the RoCSP algorithm and handles the non-stationarity as uncertainty using the interval type-2 fuzzy sets in the antecedent of fuzzy rules. A five layered modified Takagi-Sugeno-Kang interval type-2 fuzzy inference mechanism forms the structure and the learning algorithm uses a self-regulatory mechanism. Further, the SRIT2NFIS classifier is used to find the desired spectral filters by eliminating those frequency bands that do not affect the classification performance. The performance of the proposed system has been evaluated using two publicly available BCI competition data sets and compared with other existing algorithms like FBCSP, DFBCSP and BSSFO. The results indicate improved performances of the proposed algorithm. Finally, the proposed system is employed to control the movement of a quadcopter.

BACKGROUND: Protein secondary structure prediction (SSP) has been an area of intense research interest. Despite advances in recent methods conducted on large datasets, the estimated upper limit accuracy is yet to be reached. Since the predictions of SSP methods are applied as input to higher-level structure prediction pipelines, even small errors may have large perturbations in final models. Previous works relied on cross validation as an estimate of classifier accuracy. However, training on large numbers of protein chains compromises the classifier ability to generalize to new sequences. This prompts a novel approach to training and an investigation into the possible structural factors that lead to poor predictions. Here, a small group of 55 proteins termed the compact model is selected from the CB513 dataset using a heuristics-based approach. In a prior work, all sequences were represented as probability matrices of residues adopting each of Helix, Sheet and Coil states, based on energy calculations using the C-Alpha, C-Beta, Side-chain (CABS) algorithm. The functional relationship between the conformational energies computed with CABS force-field and residue states is approximated using a classifier termed the Fully Complex-valued Relaxation Network (FCRN). The FCRN is trained with the compact model proteins. RESULTS: The performance of the compact model is compared with traditional cross-validated accuracies and blind-tested on a dataset of G Switch proteins, obtaining accuracies of ∼81 %. The model demonstrates better results when compared to several techniques in the literature. A comparative case study of the worst performing chain identifies hydrogen bond contacts that lead to Coil ⇔ Sheet misclassifications. Overall, mispredicted Coil residues have a higher propensity to participate in backbone hydrogen bonding than correctly predicted Coils. CONCLUSIONS: The implications of these findings are: (i) the choice of training proteins is important in preserving the generalization of a classifier to predict new sequences accurately and (ii) SSP techniques sensitive in distinguishing between backbone hydrogen bonding and side-chain or water-mediated hydrogen bonding might be needed in the reduction of Coil ⇔ Sheet misclassifications.

This paper addresses the key problems of nonstationarity and influence of artifacts in electroencephalogram (EEG)-based brain-computer-interface (BCI) systems. The nonstationary nature of EEG data arises due to the physiological/instrumental differences in intra/intersession of the data generation process. This paper proposes a robust common spatial pattern feature extraction algorithm (RoCSP) to overcome the effects of artifacts and a self-regulated interval type-2 neuro-fuzzy inference system (SRIT2NFIS) to handle this inherent nonstationarity. Combined together, this approach is referred to as RoCSP-SRIT2NFIS. The RoCSP algorithm provides better features than the CSP algorithm by excluding those trials that are affected by the artifacts. SRIT2NFIS uses the features generated by the RoCSP algorithm as input and handles the nonstationarity as an uncertainty using the interval type-2 fuzzy sets in the antecedent of fuzzy rules. A self-regulatory learning mechanism is used to evolve the structure automatically and learn the parameters of the network. A regularized projection-based learning algorithm and a modified rule addition criterion are also proposed to improve the generalization performance of SRIT2NFIS. Using benchmark datasets, performance evaluation has been carried out, and the results indicate that, compared with other existing algorithms, RoCSP-SRIT2NFIS produces higher classification accuracy of 3-5% in simple tasks like left-right classification and 6-8% in complex tasks like foot-tongue classification. In addition, a statistical analysis of the performance results indicates that RoCSP-SRIT2NFIS performs better and is more suitable for an efficient BCI.

Humans seek to select the best decision for a given problem in a process that is highly efficient and often ends with success. This is due to a high-order thinking skill: metacognition, which enables humans to be successful decision makers by constantly monitoring their cognitive activities based on earlier experience. Besides this, the social aspect of metacognition helps humans in monitoring their cognitive activities based on their peers experience and knowledge. Inspired by this, we propose HumanCog: a generic 3-layer architecture for solving optimization problems. HumanCog functions in a way that mimics human cognitive and metacognitive (self as well as social) behavior. The three layers in the network are cognitive layer, metacognitive layer and social cognitive layer. These three layers interact with each other such that accurate decision is made. As an initial work, we provide a simple realization of the HumanCog referred to as HumanCog-ver1, which self-regulates decision based on best experience. The performance evaluation on CEC 2015 and 2013 benchmark problems indicates promising results.

This paper presents a new particle swarm optimization (PSO) algorithm incorporating the concept of mentoring in the learning process for finding the optimum solution, referred to as a Mentoring based Particle Swarm Optimization (MePSO) algorithm. In human learning principles, mentoring provides both self and social cognizance through guidance, direction and momentum for the learners. Such a mentoring concept is integrated in PSO for faster convergence where all the particles are divided into dynamically changing three groups, namely, mentors, mentees and independent learners. In a given iteration, the particles classified as the mentees are supported by the mentor particles for searching in the potential region and eventually diverting the search towards the optimum solution. All the remaining particles perform independent search by employing social-awareness of the global search direction for intelligent exploration of the solution space. The proposed MePSO algorithm has been evaluated using 15 benchmark test functions from CEC2013 and the performance has been compared with different variants of PSO algorithms reported in the literature. The comparative results and the statistical analysis clearly indicate that MePSO performs better with faster convergence characteristics.

This paper presents an improved version of the recently proposed Self Regulating Particle Swarm Optimization (SRPSO) algorithm referred to as improved Self Regulating Particle Swarm Optimization (iSRPSO) algorithm. In the iSRPSO algorithm, the last two least performing particles are observed with different perception and they adopt a different learning strategy for velocity update. These particles get a directional update from the best particle and the next top three better performing particles for divergence of their search directions towards better solutions. This provides direction and momentum to these least performing particles and enhances their awareness of the search space. Performance of iSRPSO has been compared with SRPSO on a unimodal and a multimodal benchmark function from CEC2005 where a significant performance improvement closer to the optimum solution has been observed. Further, the performance of iSRPSO has been investigated using both the 10D and 30D CEC2015 bound constrained single-objective computationally expensive numerical optimization problems. The performance of iSRPSO on 10D problems have been compared with both the PSO and SRPSO algorithms where the solutions of iSRPSO are closer to the true optimum value compared to the other two algorithms.

The Common Spatial Pattern (CSP) is an effective algorithm used in EEG based Brain Computer Interface (BCI) to extract discriminative features, however, its effectiveness depends upon the subject-specific frequency bands. Also, the generated features using CSP are non-stationary in nature. In this paper, we propose a Meta-cognitive Interval type-2 Neuro-Fuzzy Inference System to handle non-stationarity in CSP features with recursive band elimination to find subject-specific frequency bands, together known as (McIT2NFIS-RBE). McIT2NFIS uses the non-stationary features generated by CSP as its input and models it as uncertainty using Interval type-2 fuzzy sets in the antecedent of fuzzy rules. The recursive band elimination (RBE) employs the McIT2NFIS training algorithm to recursively eliminate all the features of a band, one at a time. It aims to improve the performance by removing features of a band one at a time, whose elimination will not have any effect on the training performance. The performance of McIT2NFIS-RBE is evaluated using the publicly available dataset-IIa from BCI competition dataset IV [26]. The results highlight the performance of McIT2NFIS-RBE over other algorithms.

In this paper, an automatic seizure detection technique using multichannel EEG is proposed based on Metacognitive Complex-valued Interval Type-2 Fuzzy Inference System (McCIT2FIS). A wavelet chaos theory based feature extraction is employed to extract the features from EEG signal as it can handle the non stationarity in data and Sparse Multinomial Logistic Regression via Bayesian L1 Regularisation (SBMLR) based feature selection is employed to select the most discriminative features. McCIT2FIS is employed to classify the samples as either interictal or ictal EEG segment as it has been shown to be capable of handling noisy data by virtue of Interval Type-2 fuzzy sets, and is good at classification because of its ability to handle complex-valued data. Further, we have also shown that the feature selected using SBMLR can be successfully mapped back to the channels allowing us to identify the epileptogenic regions of the brain. The performance of the McCIT2FIS was also compared with the support vector machines and the results indicate that McCIT2FIS is better capable of detecting seizure based on EEG signals.

Interval Type-2 fuzzy systems have been shown to be extremely capable of handling vagueness as well as uncertainty in data, while complex-valued fuzzy sets have been demonstrated to be capable of solving classification problems efficiently. This paper combines their collective advantage to propose a complex-valued Interval Type-2 Fuzzy Inference System (referred to as CIT2FIS). To derive the fuzzy rules, a Recursive Least Squares based algorithm is proposed. The proposed algorithm evolves (add/ prune) and adapts the rules in an evolving online fashion. During sequential learning, the network monitors the error and knowledge contained in the current sample and either rules are evolved (added, pruned) to capture the knowledge in the sample, or the rule parameter updated. Upon rule addition, the centers are determined based on the current input and the output weights are analytically determined such that a least squares fit is obtained. This ensures that the rule retain its interpretability and accuracy. Parameter update is based on recursive least squares based approach. In order to maintain the parsimony of the network, a data-driven rule pruning scheme is employed. To further enhance the generalization ability of the network, wellknown meta-cognitive learning mechanism is employed in this work. The performance of the proposed CIT2FIS is evaluated on a set of real-valued classification problems. The performance comparison with other state-of-the-art complex-valued as well as fuzzy classifiers clearly highlights the advantage of the proposed work.

This paper presents a two stage learning algorithm for a Growing-Pruning Spiking Neural Network (GPSNN) for pattern classification problems. The GPSNN uses three layered network architecture with input layer employing a modified population coding and, leaky integrate-and-fire spiking neurons in the hidden and output layers. The class label for a sample is determined according to the output neuron with minimum spike latency. The learning algorithm for the GPSNN employs a two stage learning mechanism. In the first stage, the hidden layer is grown and adapted to map the inputs to a hyperdimensional space. In the second stage, the hidden layer neurons with low dominance are pruned and the response of the most dominant neurons is mapped to the output space. The proposed approach has been evaluated on benchmark data sets from the UCI machine learning repository and the results were compared with batch as well as online spiking neural networks. The results clearly highlight that the GPSNN can achieve better performances using a compact network structure.

This paper presents an approach for automatic diagnosis of Autism Spectrum Disorder (ASD) among males using functional Magnetic Resonance Imaging (fMRI). fMRI has the capability to identify any abnormal neural interactions that may be responsible for behavioral symptoms observed in ASD patients. In this paper, the regional homogeneity of the voxels in the 116 regions of the automated anatomical labeling (AAL) atlas of the brain are used as features which result in a large set of 54837 features. Chi-square feature selection method is then used to identify the most significant features and only these features are then used for classification with a metacognitive radial basis function classifier. Since genetic studies have indicated that ASD manifests differently in males and females, a large scale study specific to males is highlighted here using the publicly available preprocessed fMRI dataset from the Autism Brain Imaging Data Exchange (ABIDE), unlike existing studies which are either smaller in scale or consider both males and females together. Among the males, it is shown here that the classification performance can be improved (by up to 10%) by considering adults and adolescents separately. By using Chi-square algorithm the number of features was reduced drastically to lower than 200 in contrast to the thousands of features that have been used in recent studies.

This paper presents an automatic, non-invasive method for detecting Autism Spectrum Disorders (ASD) among males using structural Magnetic Resonance Imaging. Whole brain Voxel Based Morphometry (VBM) analysis is first used to identify the brain regions that are affected for ASD patients and gray matter probability in these regions are used as features for classification. In contrast to existing studies which are small in scale, this paper presents a large scale study using the publicly available dataset from Autism Brain Imaging Data Exchange. Taking a cue from genetic studies which indicate that ASD manifests differently among males and females, this paper considers males only. Even among males, this study shows that better classification accuracy can be achieved by considering adult and adolescent males separately. By using a Metacognitive Radial basis Function Network classifier, classification accuracy of 59.73%, 61.49% and 70.41% is achieved when considering all males, adolescent males and adult males respectively. This is about (5 to 10%) higher than Support Vector Machine classifier which is commonly used in literature for this problem and about (6 to 15%) higher than Naive Bayes classifier. It is also found that all three classifiers perform better when considering adult and adolescent males separately instead of considering all males together underscoring the need to consider different age-groups separately for ASD detection. VBM analysis indicates that the precentral gyrus, motor cortex, medial frontal gyrus and the paracentral lobule areas are possibly affected for adolescent males with ASD while the superior frontal gyrus and the frontal eye fields areas are possibly affected for adult males with ASD.

The use of large number of channels in EEG based Motor-imagery Brain Computer Interfaces (BCI) may cause long preparation time and redundancy of data. In this paper, we propose a Cohen's d effect-size based channel selection algorithm which eliminates the redundant channels while improving the classification performance. This method (referred to as Effect-size based CSP (E-CSP)) eliminates the channels that do not carry information that distinguishes the two tasks. First, it removes the noisy trials for a channel followed by Cohen's d based effect-size calculation to determine the redundant channels. Using two publicly available BCI competition data sets, the performance of E-CSP algorithm is compared with other existing algorithms like CSP and SCSP. Results indicate that the E-CSP algorithm produces a higher classification accuracy compared to the other algorithms using lesser number of channels in a non-iterative manner.

BACKGROUND: Traditional cancer treatments have centered on cytotoxic drugs and general purpose chemotherapy that may not be tailored to treat specific cancers. Identification of molecular markers that are related to different types of cancers might lead to discovery of drugs that are patient and disease specific. This study aims to use microarray gene expression cancer data to identify biomarkers that are indicative of different types of cancers. Our aim is to provide a multi-class cancer classifier that can simultaneously differentiate between cancers and identify type-specific biomarkers, through the application of the Binary Coded Genetic Algorithm (BCGA) and a neural network based Extreme Learning Machine (ELM) algorithm. RESULTS: BCGA and ELM are combined and used to select a subset of genes that are present in the Global Cancer Mapping (GCM) data set. This set of candidate genes contains over 52 biomarkers that are related to multiple cancers, according to the literature. They include APOA1, VEGFC, YWHAZ, B2M, EIF2S1, CCR9 and many other genes that have been associated with the hallmarks of cancer. BCGA-ELM is tested on several cancer data sets and the results are compared to other classification methods. BCGA-ELM compares or exceeds other algorithms in terms of accuracy. We were also able to show that over 50% of genes selected by BCGA-ELM on GCM data are cancer related biomarkers. CONCLUSIONS: We were able to simultaneously differentiate between 14 different types of cancers, using only 92 genes, to achieve a multi-class classification accuracy of 95.4% which is between 21.6% and 38% higher than other results in the literature for multi-class cancer classification. Our findings suggest that computational algorithms such as BCGA-ELM can facilitate biomarker-driven integrated cancer research that can lead to a detailed understanding of the complexities of cancer.

In this paper, we propose an evolving interval type-2 neurofuzzy inference system (IT2FIS) and its fully sequential learning algorithm. IT2FIS employs interval type-2 fuzzy sets in the antecedent part of each rule and the consequent realizes Takagi-Sugeno-Kang fuzzy inference mechanism. In order to render the inference fast and accurate, we propose a data-driven interval-reduction approach to convert interval type-1 fuzzy set in antecedent to type-1 fuzzy number in the consequent. During learning, the sequential algorithm learns a sample one-by-one and only once. The IT2FIS structure evolves automatically and adapts its network parameters using metacognitive learning mechanism concurrently. The metacognitive learning regulates the learning process by appropriate selection of learning strategies and helps the proposed IT2FIS to approximate the input-output relationship efficiently. An evolving IT2FIS employing a metacognitive learning algorithm is referred to as McTI2FIS. Performance of metacognitive interval type-2 neurofuzzy inference system (McIT2FIS) is evaluated using a set of benchmark time-series problems and is compared with existing type-2 and type-1 fuzzy inference systems. Finally, the performance of the proposed McIT2FIS has been evaluated using a practical stock price-tracking problem. The results clearly highlight that McIT2FIS performs better than other existing results in the literature.

Attention Deficiency Hyperactivity Disorder (ADHD) as a disruptive behavior disorder is receiving lots of attention because of its complexity and need for early detection. This paper presents a study on identification of potential biomarkers in the diagnosis of ADHD based on the structural-MRI of the brain obtained through ADHD-200 competition data set. The region of the brain considered here is "hippocampus". The grey matter probability of the T1 images is segmented followed by tissue alignment and inter subject normalization. Then, the voxels of the hippocampus are segregated using a region-of-interest mask, and the grey matter tissue probability values are obtained. These values are then used as features to classify ADHD patients against typically developing controls using a projection based learning algorithm for a meta-cognitive radial basis function network (PBL-McRBFN) and compared the results with that of support vector machines. Initially we take all the voxels of hippocampus for our study and then we have selected the most relevant voxels as a biomarker using Chi-square approach and developed a classifier to diagnosis ADHD. The results clearly highlight that use of hippocampus from the structural-MRI is sufficient to diagnosis ADHD to certain degree of confidence.

In this paper, a fast learning methodology for neuro-fuzzy inference system (NFIS) referred to as meta-cognitive fuzzy extreme learning machine (McFELM) is proposed. It is based on the original OS-Fuzzy-ELM algorithm incorporating principles of human meta-cognition to make the learning more effective. McFELM has two components: the cognitive component and the meta-cognitive component. The cognitive component is a fuzzy extreme learning machine which learn sequential data in a one-by-one mode or a chunk-by-chunk mode with fixed or varying chunk size, while the meta-cognitive component controls the learning process of the cognitive component using a self-regulating mechanism to decide what-to-learn, when-to-learn, and how-to-learn. Unlike the OS-Fuzzy-ELM algorithm which uses all arriving samples to update the output weight matrix, the proposed algorithm employs different strategies namely sample deletion, sample reserve and sample learning strategy to decide whether the data will be deleted directly, reserved for later use or used immediately. Instantaneous error is used to select the best learning strategy. The evaluation of McFELM is presented doing simulations on a nonlinear system identification problem and a set of benchmark regression problems from UCI machine leaning repository. The results show that the proposed McFELM produces better performance compared with existing algorithms.

In this paper, a computationally efficient Interval Type-2 Neuro-Fuzzy Inference System (IT2FIS) and its Meta-Cognitive projection based learning (PBL) algorithm is presented, together referred as PBL-McIT2FIS. A six layered network with computationally cheap type-reduction technique is proposed, rendering the inference mechanism faster. During learning, the projection based learning algorithm assumes that IT2FIS has no rules in the beginning, and the learning algorithm adds rules to the network and updates it depending on the prediction error and relative knowledge present in the current sample. As each sample is presented to the network, the meta-cognitive component of the learning algorithm decides what-to-learn, when-to-learn and how-to-learn it, depending on the instantaneous error and spherical potential of the current sample. Whenever a new rule is added or an existing rule is updated, a projection based learning algorithm computes the optimal output weights by minimizing the total error in the network in a computationally efficient manner. The performance of PBL-McIT2FIS is evaluated on a set of benchmark problem and compared to other state-of-the-art algorithms available in literature. The results indicate superior performance of PBL-McIT2FIS.

In this paper, we develop a new sequential learning algorithm for a spiking neural network classifier. The algorithm handles the input features that are not in the form of a spike train but in a real-valued (analog) form. The sequential learning algorithm evolves the number of spiking neuron automatically based on the information present in the current sample and results in a compact architecture. Hence, it is referred to as a Minimal Spiking Neural Network (MSNN). The learning algorithm can either add a new neuron to the network or update the parameters of the existing neurons based on the information contained in the arriving samples. The update rule uses excitatory/inhibitatory rule to capture the knowledge contained in the current sample. Performance evaluation of the proposed MSNN is presented using two benchmark problems from the UCI machine learning repository, namely, the Iris flower classification and Wisconsin breast cancer problem and the results are compared with other existing spiking neural algorithms like SpikeProp, MuSpiNN and Multi-spike learning algorithms. The results clearly indicate the better performance of MSNN with a compact architecture.

This paper presents a complex-valued interval type-2 neuro-fuzzy inference system (CIT2FIS) and derive its metacognitive projection-based learning (PBL) algorithm. Metacognitive CIT2FIS (Mc-CIT2FIS) consists of a CIT2FIS, which realizes Takagi-Sugeno-Kang type inference mechanism, as its cognitive component. A PBL with self-regulation is its metacognitive component. The rules of CIT2FIS employ interval type-\(2~q\) -Gaussian membership functions that can represent different radial basis functions for different values of \(q\) . As each sample is presented to the network, the metacognitive component monitors the hinge-loss error and class-specific knowledge potential of the current sample to efficiently decide on what-to-learn, when-to-learn, and how-to-learn it. When a new rule is added or existing rules are updated, the optimal parameters of CIT2FIS corresponding to the minimum of the hinge-loss error function are computed using a PBL algorithm derived using the Wirtinger calculus. The performance of Mc-CIT2FIS is evaluated on a set of benchmark real-valued classification problems from the UCI machine learning repository. A circular transformation is used to convert the real-valued features to the complex-valued features in these problems. The performance comparison and statistical study clearly show the superior classification ability of Mc-CIT2FIS. Finally, the proposed complex-valued network is used to solve a practical human action recognition problem that is represented by complex-valued optical flow-based feature set, and a human emotion recognition problem represented using complex-valued Gabor filter-based features. The performance results on these problems substantiate the superior classification ability of Mc-CIT2FIS.

Structures with fixed-base will produce high accelerations and inter-storey drifts and move laterally during earthquake. The presence of base isolation devices between ground and the structure, reduces the structural vibrations. To maintain the seismic response of structures within safety, service and comfort limits, the combination of base isolators and feedback controllers have been proposed in recent years. This paper proposes a discrete direct adaptive extreme learning machine (ELM) controller for the active control of non linear base isolated building with hysteretic isolation system. The controller is constructed based on a single hidden layer feed forward network and the parameters of the network are adapted using extreme learning machine (ELM) algorithm. In this work, different from the original ELM algorithm the output weights of the network are updated using Lyapunov stability approach, to guarantee the stability of the structure. The performance of the proposed controller is verified on a non-linear three dimensional benchmark base-isolated structure by exciting the structure with three earthquake samples. The result shows that the proposed controller is effective in reducing the seismic responses of the isolation members as well as the superstructure.

This paper presents a neural adaptive flight controller for ducted fan UAVs which are capable of vertical takeoff and landing(VTOL). These ducted fan propulsion systems pose great challenges in aerodynamics and control and we propose a backstepping neural adaptive control law to track its nonlinear dynamics. This controller can handle unmodeled dynamics and external disturbances as well, providing stability to the vehicle. A single layer radial basis neural network is used to approximate the unmodeled dynamics and vehicle stability is guaranteed through Lyapunov synthesis. For simulation study six degree of freedom model (6-DOF) is implemented in MATLAB along with the proposed control approach. The performance of the controller is evaluated using nonlinear bop-up maneuver and the necessary stability and tracking performance of the UAV have been investigated.

In this paper, we propose a Projection Based Learning (PBL) algorithm for a Meta-Cognitive Neuro-Fuzzy Inference (McFIS) together referred to as PBL-McFIS. McFIS consists of a cognitive component, which is a zero-order Takagi-Sugeno-Kang adaptive neuro-fuzzy inference system, and a meta-cognitive component, which is a self-regulatory learning mechanism for the neuro-fuzzy inference system. The learning in the cognitive component begins with zero rules, and as new samples are presented to the network, the meta-cognitive component monitors the hinge-loss error and spherical potential of the current sample to efficiently decide on what-to-learn, when-to-learn and how-to-learn. In this work we employ PBL-McFIS to solve classification problems and hence the monitory signals employ class-specific self-adaptive thresholds to decide on efficient learning strategies. These thresholds are self-adapted such that the trained network is compact and avoids over-fitting. During addition of new rules or updating of existing rules, the optimal output weights corresponding to the minimum hinge-loss error is computed using PBL algorithm. The learning algorithm considers class-specific as well as class overlap factors during training. The performance of PBL-McFIS is evaluated on a set of benchmark classification problems. The statistical performance analysis with other state-of-the-art neuro-fuzzy inference systems and SVM indicate improved classification ability of the proposed algorithm.

This paper proposes a novel sequential learning algorithm based on zero-density maximization and extended Kalman filter, together referred to as ZDM-EKF, for a Takagi-Sugeno-Kang fuzzy inference system. The sequential learning begins with zero rules, and rules are added/ pruned or updated based on the knowledge contained in the network and prediction error of the current sample. As each sample is presented to the network, the network monitors the spherical potential and mean-squared error and either adds a new rule or updates the parameters of the nearest rules employing an extended Kalman filtering scheme. The Kalman filter estimates the optimal network parameter based on maximizing the error density at origin. This results in a simple and efficient cost function with better ability to learn higher-order statistical behavior in comparison to error based cost function. The performance of the proposed ZDM-EKF based learning algorithm is evaluated on a set of four synthetic function approximation as well as time-series prediction problems. The performance analysis indicates superior performance of the proposed algorithm.

In this paper, we present a novel approach for classification of Mild Cognitive Impairment (MCI) and normal subjects from Magnetic Resonance Images (MRI) using a proposed `sequential Projection Based Learning for Meta-cognitive q-Gaussian Radial Basis Function Network (PBL-McqRBFN)' classifier. The McqRBFN has two components, namely, a cognitive component and a meta-cognitive components. The cognitive component is a single hidden layer Radial Basis Function (RBF) network with a q-Gaussian activation function, that allows different RBF's in one network, like the Gaussian, the Inverse Multiquadratic, and the Cauchy functions, by changing a real q-parameter. The meta-cognitive component present in McqRBFN helps in selecting proper samples to learn based on its current knowledge and evolve architecture automatically. The McqRBFN employs a sequential Projection Based Learning (PBL) algorithm to reduce the computational effort used in training. For simulation studies, we have used MRI data from the Alzheimer's Disease Neuroimaging Initiative database. Voxel Based Morphometry (VBM) is used for feature extraction from MRI data and extracted VBM features are fed into the PBL-McqRBFN classifier. The experimental results show that our proposed PBL-McqRBFN classifier can accurately differentiate MCI and normal subjects.

Knowledge of the various protein functions is essential to understand the manifestation of diseases and develop suitable drugs to alleviate them. As proteins form conformational patterns like α-helix and β-strands that eventually fold up into 3-D structure, prediction of the secondary structure of proteins is an important intermediate step in understanding the final structure of proteins that are vital for performing biological functions. Thus, there is a need to predict the secondary structure of proteins accurately and efficiently. Recent studies in machine learning have shown that complex-valued neural networks have better decision making ability than real-valued networks. Therefore, we use a Fully Complex-valued Relaxation Network (FCRN) classifier to predict the secondary structure of proteins in this paper. FCRN classifier is a single hidden layer neural network classifier with nonlinear input, hidden and output layers. The neurons in the input layer convert the real-valued input features to the Complex domain using a circular transformation. The neurons in the hidden layer employ a fully complex-valued sech activation function and those in the output layer employ the fully complex-valued exp activation function. For constant random input parameters, FCRN estimates the output weights corresponding to the minimum energy point of a logarithmic error function that represents both the magnitude and phase error explicitly. The prediction performance of FCRN is compared against the best results available in the literature for this problem. Our results show that FCRN presents higher or comparable prediction accuracy than other classifiers available in the literature.

This paper presents an evolving spiking neural network namely, `Basis Coupled Evolving Spiking Neural Network (BCESNN)' and its learning algorithm to solve real-valued pattern recognition problems. BCESNN is a two-layered neuron model with afferent neurons in the input layer and efferent neurons in the output layer. The afferent neurons in the input layer convert the real-valued input feature to a train of spikes using a bank of Gaussian Receptive Field (GRF) for each individual feature. The number of GRF per feature is fixed a priori. Each efferent neuron in the output layer is associated to a class. Efferent neurons are integrate-and-fire type neuron. BCESNN has an evolving architecture that uses basis coupled rank order learning (BCROL) algorithm to estimate the number of output neurons and the network parameters. Each sample is presented only once to the network. When a new sample is presented to the network either we add a neuron or we update an existing neuron. Weight estimation for added neuron is done using BCROL and weight update is done using Euclidean distance based distance measure. In the performance section we conducted three different experiments. Firstly we compared the performance of BCROL against Rank Order Learning(ROL). Next, we evaluated the performance of BCESNN on benchmark classification problems from the UCI machine learning repository. Finally, we evaluated the performance of a sparsely connected BCESNN against fully connected BCESNN where connectivity refers to the number of GRF connected to the afferent neurons.

In this paper, we present an approach for the diagnosis of Autism Spectrum Disorder (ASD) from Magnetic Resonance Imaging (MRI) scans with Voxel-Based Morphometry (VBM) detected features using Projection Based Learning (PBL) algorithm for a Meta-cognitive Radial Basis Function Network (McRBFN) classifier. McRBFN emulates human-like meta-cognitive learning principles. As each sample is presented to the network, the McRBFN uses the estimated class label, the maximum hinge error and class-wise significance to address the self-regulating principles of what-to-learn, when-to-learn and how-to-learn in a meta-cognitive framework. Initially, McRBFN begins with zero hidden neurons and adds required number of neurons to approximate the decision surface. When a neuron is added, its parameters are initialized based on the sample overlapping conditions. The output weights are updated using a PBL algorithm such that the network finds the minimum point of an energy function defined by the hinge-loss error. Moreover, as samples with similar information are deleted, over-training is avoided. The PBL algorithm helps to reduce the computational effort used in training. For simulation studies, we have used MR images from the Autism Brain Imaging Data Exchange (ABIDE) data set. The performance of the PBL-McRBFN classifier is evaluated on complete morphometric features set obtained from the VBM analysis. The performance evaluation study clearly indicates the superior performance of PBL-McRBFN classifier over other classification algorithms.

In this paper, we present a metacognitive sequential learning algorithm for a neuro-fuzzy inference system for classification tasks, which is referred to as a “metacognitive neuro-fuzzy inference system (McFIS).” The McFIS learning algorithm is developed based on the principles of the best human learning strategy, viz., a self-regulatory learning strategy in a metacognitive framework. McFIS has two components: a cognitive component and a metacognitive component. A neuro-fuzzy inference system forms the cognitive component of the McFIS, and a self-regulatory learning mechanism forms its metacognitive component. The learning ability of the cognitive component is monitored and controlled by the self-regulatory learning mechanism. For each sample in the training dataset, the metacognitive component uses its self-adaptive thresholds to choose one of the following learning strategies based on the criteria that depends on class-specific knowledge: 1) sample deletion; 2) sample learning; and 3) sample reserve. Thus, the metacognitive component decides what-to-learn, when-to-learn, and how-to-learn the training samples. When a new rule is added, the parameters of the new rule are assigned such that the rule has minimum overlapping with the adjacent rules as well as the localization property of the Gaussian rules is efficiently exploited. Performance of the McFIS is evaluated using several well-known benchmark multicategory/binary classification datasets from the University of California, Irvine machine learning repository and on a practical human action recognition problem. The results clearly indicate that the proposed metacognitive learning helps the McFIS achieve better performance than other existing classifiers.

In this paper, we present a sequential projection-based metacognitive learning algorithm in a radial basis function network (PBL-McRBFN) for classification problems. The algorithm is inspired by human metacognitive learning principles and has two components: a cognitive component and a metacognitive component. The cognitive component is a single-hidden-layer radial basis function network with evolving architecture. The metacognitive component controls the learning process in the cognitive component by choosing the best learning strategy for the current sample and adapts the learning strategies by implementing self-regulation. In addition, sample overlapping conditions and past knowledge of the samples in the form of pseudosamples are used for proper initialization of new hidden neurons to minimize the misclassification. The parameter update strategy uses projection-based direct minimization of hinge loss error. The interaction of the cognitive component and the metacognitive component addresses the what-to-learn, when-to-learn, and how-to-learn human learning principles efficiently. The performance of the PBL-McRBFN is evaluated using a set of benchmark classification problems from the University of California Irvine machine learning repository. The statistical performance evaluation on these problems proves the superior performance of the PBL-McRBFN classifier over results reported in the literature. Also, we evaluate the performance of the proposed algorithm on a practical Alzheimer's disease detection problem. The performance results on open access series of imaging studies and Alzheimer's disease neuroimaging initiative datasets, which are obtained from different demographic regions, clearly show that PBL-McRBFN can handle a problem with change in distribution.

This paper presents a fully complex-valued relaxation network (FCRN) with its projection-based learning algorithm. The FCRN is a single hidden layer network with a Gaussian-like sech activation function in the hidden layer and an exponential activation function in the output layer. For a given number of hidden neurons, the input weights are assigned randomly and the output weights are estimated by minimizing a nonlinear logarithmic function (called as an energy function) which explicitly contains both the magnitude and phase errors. A projection-based learning algorithm determines the optimal output weights corresponding to the minima of the energy function by converting the nonlinear programming problem into that of solving a set of simultaneous linear algebraic equations. The resultant FCRN approximates the desired output more accurately with a lower computational effort. The classification ability of FCRN is evaluated using a set of real-valued benchmark classification problems from the University of California, Irvine machine learning repository. Here, a circular transformation is used to transform the real-valued input features to the complex domain. Next, the FCRN is used to solve three practical problems: a quadrature amplitude modulation channel equalization, an adaptive beamforming, and a mammogram classification. Performance results from this paper clearly indicate the superior classification/approximation performance of the FCRN.

In this paper, we present a novel approach with Voxel-Based Morphometry (VBM) detected features using a proposed ‘Projection Based Learning for Meta-cognitive Radial Basis Function Network (PBL-McRBFN)’ classifier for the detection of Alzheimer's disease (AD) from Magnetic Resonance Imaging (MRI) scans. McRBFN emulates human-like meta-cognitive learning principles. As each sample is presented to the network, McRBFN uses the estimated class label, the maximum hinge error and class-wise significance to address the self-regulating principles of what-to-learn, when-to-learn and how-to-learn in a meta-cognitive framework. Initially, McRBFN begins with zero hidden neurons and adds required number of neurons to approximate the decision surface. When a neuron is added, its parameters are initialized based on the sample overlapping conditions. The output weights are updated using a PBL algorithm such that the network finds the minimum point of an energy function defined by the hinge-loss error. Moreover, as samples with similar information are deleted, over-training is avoided. The PBL algorithm helps to reduce the computational effort used in training. For simulation studies, we have used well-known open access series of imaging studies data set. The performance of the PBL-McRBFN classifier is evaluated on complete morphometric features set obtained from the VBM analysis and also on reduced features sets from Independent Component Analysis (ICA). The performance evaluation study clearly indicates the superior performance of PBL-McRBFN classifier over results reported in the literature.

In this paper, we propose a ‘Meta-cognitive Radial Basis Function Network (McRBFN)’ and its ‘Projection Based Learning (PBL)’ algorithm for classification problems. McRBFN emulates human-like meta-cognitive learning principles. As each sample is presented to the network, McRBFN uses the estimated class label, the maximum hinge error and class-wise significance to address the self-regulating principles of what-to-learn, when-to-learn and how-to-learn in a meta-cognitive framework. McRBFN addresses the what-to-learn by choosing samples to participate in the learning process, also deleting samples with information similar to that already learnt by the network. A few samples that satisfy neither of these criteria are pushed to the rear end of the training data stack to be used in future, thereby satisfying the when-to-learn. The how-to-learn component of meta-cognition is addressed by using the participating samples to either add a neuron or update the output weights. Initially, McRBFN begins with zero hidden neurons and adds required number of neurons to approximate the decision surface. When a neuron is added, its parameters are initialized based on the sample overlapping conditions. The output weights are updated using a PBL algorithm such that the network finds the minimum point of an energy function defined by the hinge-loss error. The use of human meta-cognitive principles ensures efficient learning. Moreover, as samples with similar information are deleted, overtraining is avoided. The PBL algorithm helps to reduce the computational effort used in training. The performance of the PBL-McRBFN classifier is evaluated using a set of benchmark classification problems from the UCI machine learning repository. The performance evaluation study on these problems clearly indicates the superior performance of PBL-McRBFN classifier over results reported in the literature.

In this paper, we propose a Meta-Cognitive Neuro-Fuzzy Inference System (McFIS) for accurate detection of human actions from video sequences. In this paper, we employ optical flow based features as they can represent information from local pixel level to global object level between two consecutive image planes. The functional relationship between these optical flow based features and action classes is approximated using McFIS classifier. The sequential learning algorithm is developed based on the principles of self-regulation observed in human meta-cognition. McFIS decides on what-to-learn, when-to-learn and how-to-learn based on the knowledge stored in the classifier and the information contained in the new training sample. The sequential learning algorithm of McFIS is controlled and monitored by the meta-cognitive components which uses class-specific and knowledge based criteria along with self-regulatory thresholds to decide on one of the following strategies: a) sample deletion b) sample learning and c) sample reserve. Performance of proposed McFIS based human action recognition system is evaluated using benchmark Weizmann and KTH video sequences. The simulation results are compared with well known support vector machine classifier and also with state-of-the-art action recognition results reported in the literature. The results clearly indicates McFIS action recognition system achieves better performances with minimal computational effort.

In this paper, we propose a Meta-Cognitive Neuro-Fuzzy Inference System (McFIS) for recognition of emotions from facial features. Local binary patterns have been proven to effectively describe the statistical characteristics of face image as it contains information related to edges, spots, etc. The aim of McFIS is to approximate the functional relationship between the facial features and various emotions. McFIS classifier and its sequential learning algorithm is developed based on the principles of self-regulation observed in human meta-cognition. McFIS decides on what-to-learn, when-to-learn and how-to-learn based on the knowledge stored in the classifier and the information contained in the new training samples. The sequential learning algorithm of McFIS is controlled and monitored by the meta-cognitive components which uses class-specific, knowledge based criteria along with self-regulatory thresholds to decide on one of the following strategies: a) sample deletion b) sample learning and c) sample reserve. Performance of proposed McFIS based facial emotion recognition is evaluated on LBP features extracted from JAFFE database. The simulation results are compared with support vector machine classifier and other results available in literature. The results indicate the superior performance of McFIS in comparison to other algorithms.

In this paper, we present a complex-valued neuro-fuzzy inference system (CNFIS) and its gradient descent based learning algorithm developed employing Wirtinger calculus. The proposed CNFIS is a four layered network which realizes zero-order Takagi-Sugeno-Kang based fuzzy inference mechanism. CNFIS is used to predict the speed and direction of wind. Here, the speed and direction are considered as statistically independent variables and are represented as a complex-valued signal (with speed as magnitude and direction as phase). Performance of CNFIS is compared with other algorithms available in the literature and results indicate improved performance of CNFIS. The major contribution of this paper is as follows: (1) Propose a complex-valued neuro-fuzzy inference system (2) Employ Wirtinger calculus for complex-valued gradient descent algorithm (3) Solve wind speed and direction prediction problem in complex domain.

We propose a sequential Meta-Cognitive learning algorithm for Neuro-Fuzzy Inference System (McFIS) to efficiently recognize human actions from video sequence. Optical flow information between two consecutive image planes can represent actions hierarchically from local pixel level to global object level, and hence are used to describe the human action in McFIS classifier. McFIS classifier and its sequential learning algorithm is developed based on the principles of self-regulation observed in human meta-cognition. McFIS decides on what-to-learn, when-to-learn and how-to-learn based on the knowledge stored in the classifier and the information contained in the new training samples. The sequential learning algorithm of McFIS is controlled and monitored by the meta-cognitive components which uses class-specific, knowledge based criteria along with self-regulatory thresholds to decide on one of the following strategies: (i) Sample deletion (ii) Sample learning and (iii) Sample reserve. Performance of proposed McFIS based human action recognition system is evaluated using benchmark Weizmann and KTH video sequences. The simulation results are compared with well known SVM classifier and also with state-of-the-art action recognition results reported in the literature. The results clearly indicates McFIS action recognition system achieves better performances with minimal computational effort.

Recent studies on human learning reveal that self-regulated learning in a metacognitive framework is the best strategy for efficient learning. As the machine learning algorithms are inspired by the principles of human learning, one needs to incorporate the concept of metacognition to develop efficient machine learning algorithms. In this letter we present a metacognitive learning framework that controls the learning process of a fully complex-valued radial basis function network and is referred to as a metacognitive fully complex-valued radial basis function (Mc-FCRBF) network. Mc-FCRBF has two components: a cognitive component containing the FC-RBF network and a metacognitive component, which regulates the learning process of FC-RBF. In every epoch, when a sample is presented to Mc-FCRBF, the metacognitive component decides what to learn, when to learn, and how to learn based on the knowledge acquired by the FC-RBF network and the new information contained in the sample. The Mc-FCRBF learning algorithm is described in detail, and both its approximation and classification abilities are evaluated using a set of benchmark and practical problems. Performance results indicate the superior approximation and classification performance of Mc-FCRBF compared to existing methods in the literature.

Research in the field of recognizing unlimited vocabulary, online handwritten Indic words is still in its infancy. Most of the focus so far has been in the area of isolated character recognition. In the context of lexicon-free recognition of words, one of the primary issues to be addressed is that of segmentation. As a preliminary attempt, this paper proposes a novel script-independent, lexicon-free method for segmenting online handwritten words to their constituent symbols. Feedback strategies, inspired from neuroscience studies, are proposed for improving the segmentation. The segmentation strategy has been tested on an exhaustive set of 10000 Tamil words collected from a large number of writers. The results show that better segmentation improves the overall recognition performance of the handwriting system.

In this paper, we present a fast learning neural network classifier for human action recognition. The proposed classifier is a fully complex-valued neural network with a single hidden layer. The neurons in the hidden layer employ the fully complex-valued hyperbolic secant as an activation function. The parameters of the hidden layer are chosen randomly and the output weights are estimated analytically as a minimum norm least square solution to a set of linear equations. The fast leaning fully complex-valued neural classifier is used for recognizing human actions accurately. Optical flow-based features extracted from the video sequences are utilized to recognize 10 different human actions. The feature vectors are computationally simple first order statistics of the optical flow vectors, obtained from coarse to fine rectangular patches centered around the object. The results indicate the superior performance of the complex-valued neural classifier for action recognition. The superior performance of the complex neural network for action recognition stems from the fact that motion, by nature, consists of two components, one along each of the axes.

This paper presents a new approach using Voxel-Based Morphometry (VBM) detected features with a Self-adaptive Resource Allocation Network (SRAN) classifier for the detection of Alzheimer's Disease (AD) from Magnetic Resonance Imaging (MRI) scans. For feature reduction, Principal Component Analysis (PCA) has been performed on the morphometric features obtained from the VBM analysis and these reduced features are then used as input to the SRAN classifier. In our study, the MRI volumes of 30 ‘mild AD to moderate AD’ patients and 30 normal persons from the well-known Open Access Series of Imaging Studies (OASIS) data set have been used. The results indicate that the SRAN classifier produces a mean testing efficiency of 91.18% with only 20 PCA reduced features whereas, the Support Vector Machine (SVM) produces a mean testing efficiency of 90.57% using 45 PCA reduced features. Also, the results show that the SRAN classifier avoids over-training by minimizing the number of samples used for training and provides a better generalization performance compared to the SVM classifier. The study clearly indicates that our proposed approach of PCA-SRAN classifier performs accurate classification of AD subjects using reduced morphometric features.

This paper presents a fast learning fully complex-valued classifier to solve real-valued classification problems, called the `Fast Learning Complex-valued Neural Classifier' (FLCNC). The FLCNC is a single hidden layer network with a non-linear, real to complex transformed input layer, a hidden layer with a fully complex activation function and a linear output layer. The neurons in the input layer convert the real-valued input features to the Complex domain using an unique non-linear transformation. At the hidden layer, the complex-valued transformed input features are mapped onto a higher dimensional Complex plane using a fully complex-valued activation function of the type of `sech'. The parameters of the input and hidden neurons of the FLCNC are chosen randomly and the output parameters are estimated analytically which makes the FLCNC to perform fast classification. Moreover, the unique nonlinear input transformation and the orthogonal decision boundaries of the complex-valued neural network help the FLCNC to perform accurate classification. Performance of the FLCNC is demonstrated using a set of multi-category and binary real valued classification problems with both balanced and unbalanced data sets from the UCI machine learning repository. Performance comparison with existing complex-valued and real-valued classifiers show the superior classification performance of the FLCNC.

A neuro-fuzzy classifier based on the meta-cognitive principle of human self-regulated learning (Mc-FIS) is proposed in this paper. The network decides what-to-learn, when-to-learn and how-to-learn based on the current information present in the classifier and the new information present in the sample. The classifier utilizes self-regulating error based criterion to decide which sample to learn and when to learn. A rule is pruned if its significance is below a particular threshold, based on class specific information. This results in a compact network and sample deletion helps overfitting. Class specific information is used in executing the above tasks. The algorithm is evaluated on balanced and unbalanced benchmark problems from UCI machine learning repository. The results clearly indicate the superiority of the developed algorithm.

This paper presents a fast learning algorithm for a single hidden layer complex-valued neural network named as the “Fully Complex-valued Relaxation Network (FCRN)”. FCRN employs a fully complex-valued Gaussian like activation function (sech) in the hidden layer and an exponential activation function in the output layer. FCRN estimates the minimum energy state of a logarithmic error function which represents both the magnitude and phase errors explicitly to compute the optimum output weights for randomly chosen hidden layer parameters. As the weights are computed by the inversion of a nonsingular matrix, FCRN requires lesser computational effort during training. Performance studies using a synthetic function approximation problem and a QAM equalization problem show improved approximation ability of the proposed FCRN network.

A combination of Integer-Coded Genetic Algorithm (ICGA) and Particle Swarm Optimization (PSO), coupled with the neural-network-based Extreme Learning Machine (ELM), is used for gene selection and cancer classification. ICGA is used with PSO-ELM to select an optimal set of genes, which is then used to build a classifier to develop an algorithm (ICGA_PSO_ELM) that can handle sparse data and sample imbalance. We evaluate the performance of ICGA-PSO-ELM and compare our results with existing methods in the literature. An investigation into the functions of the selected genes, using a systems biology approach, revealed that many of the identified genes are involved in cell signaling and proliferation. An analysis of these gene sets shows a larger representation of genes that encode secreted proteins than found in randomly selected gene sets. Secreted proteins constitute a major means by which cells interact with their surroundings. Mounting biological evidence has identified the tumor microenvironment as a critical factor that determines tumor survival and growth. Thus, the genes identified by this study that encode secreted proteins might provide important insights to the nature of the critical biological features in the microenvironment of each tumor type that allow these cells to thrive and proliferate.

This paper presents a sequential learning algorithm for a complex-valued resource allocation network with a self-regulating scheme, referred to as complex-valued self-regulating resource allocation network (CSRAN). The self-regulating scheme in CSRAN decides what to learn, when to learn, and how to learn based on the information present in the training samples. CSRAN is a complex-valued radial basis function network with a sech activation function in the hidden layer. The network parameters are updated using a complex-valued extended Kalman filter algorithm. CSRAN starts with no hidden neuron and builds up an appropriate number of hidden neurons, resulting in a compact structure. Performance of the CSRAN is evaluated using a synthetic complex-valued function approximation problem, two real-world applications consisting of a complex quadrature amplitude modulation channel equalization, and an adaptive beam-forming problem. Since complex-valued neural networks are good decision makers, the decision-making ability of the CSRAN is compared with other complex-valued classifiers and the best performing real-valued classifier using two benchmark unbalanced classification problems from UCI machine learning repository. The approximation and classification results show that the CSRAN outperforms other existing complex-valued learning algorithms available in the literature.

An intelligent fault-tolerant control design based on nonlinear adaptive control theory is proposed for micro-satellite attitude control system, without providing for redundant actuators. An intelligent neural adaptive fault-tolerant control law, based on feedback error learning scheme is developed to aid the baseline controller. Using this control law, the control torques are suitably adjusted to achieve the desired target acquisition mode under a partial actuator failure in the roll direction. Simulation studies have been carried-out to examine the capabilities of the proposed fault-tolerant control system (***Suresh : not sure if this is right :::under misalignment of actuator and) under partial actuator failure in the roll-axis. Results clearly show that the proposed intelligent controller is robust and requires a minimal settling time during the desired target acquisition mode.

The paper describes the development of an automatic take-off system for helicopter. Design of automatic take-off control system is mandatory for manned as well as unmanned helicopter, which reduce the burden of pilot and increases the flying quality of the vehicle. In this paper, we present an It based automatic take-off system for helicopter. For this purpose, we consider typical four-bladed helicopter with conventional control. A single robust controller for altitude and attitude tracking are designed at single hovering altitude such that controller stabilized the closed loop system at all hovering altitude. The performance of the controller is evaluated in the presence of sensor noise and moderate vertical gust.

In this paper, we present an efficient learning algorithm for a Fully Complex-valued Radial Basis Function (FC-RBF) Network using a self-regulatory system. One of the important issues in gradient descent learning algorithm for complex-valued network is the proper selection of training data sequence. In general, it is assumed that the training data is uniformly distributed in the input space with non-recurrent training samples. For most real-world problems, this assumption may not be valid. Hence, one needs to develop a learning algorithm which can select proper samples for learning. This paper presents a self-regulatory system that selects samples for learning in each epoch of the batch learning scheme. The algorithm focuses on learning samples with higher errors in the same epoch, deleting samples with smaller errors from the training data set. If the samples do not satisfy both these conditions, they are neither learnt nor deleted but will be used in the next epoch for learning. As this system avoids repeated learning of similar samples, it improves the generalization performance of the FC-RBF network with a lesser computational effort. Performance studies on benchmark problems clearly show the superiority of the proposed algorithm.

This paper presents an optimization tool for launch vehicle design and trajectory optimization using bio-inspired computing algorithms and nonlinear programming. The objective is to size a launch vehicle such that the payload to lift-of-weight ratio is maximized (i.e the lift off weight is a minimum). Here, the staging problem is solved using Particle Swarm Optimization (PSO) method. With the above vehicle, an optimal trajectory is arrived at using a Real-Coded Genetic Algorithm (RCGA) and solving a nonlinear programming (NLP) by the direct shooting method. The solutions from PSO and RCGA are used for initialization of NLP variables. A case study is carried out that establishes the advantage of the proposed approach.

In this paper, a fully complex radial basis function (FC-RBF) network and a gradient descent learning algorithm are presented. Many complex-valued RBF learning algorithms have been presented in the literature using a split-complex network which uses a real activation function in the hidden layer, i.e., the activation function in these network maps Cnrarr R. Hence these algorithms do not consider the influence of phase change explicitly and hence do not approximate phase accurately. In this paper, a Gaussian like fully complex activation function sech(.) (Cnrarr C) and a well defined gradient descent learning algorithm are developed for a FC-RBF network using sech(.) as activation function. The performance evaluation of the FC-RBF network has been carried out with two synthetic complex-valued function approximation problems, a complex XOR (C-XOR) problem and a non-minimum phase equalization problem. The results indicate the better performance of the FC-RBF network compared to the existing split complex RBF network methods.

In this paper, a fully complex-valued radial basis function (FC-RBF) network with a fully complex-valued activation function has been proposed, and its complex-valued gradient descent learning algorithm has been developed. The fully complex activation function, sech(.) of the proposed network, satisfies all the properties needed for a complex-valued activation function and has Gaussian-like characteristics. It maps C(n) --> C, unlike the existing activation functions of complex-valued RBF network that maps C(n) --> R. Since the performance of the complex-RBF network depends on the number of neurons and initialization of network parameters, we propose a K-means clustering based neuron selection and center initialization scheme. First, we present a study on convergence using complex XOR problem. Next, we present a synthetic function approximation problem and the two-spiral classification problem. Finally, we present the results for two practical applications, viz., a non-minimum phase equalization and an adaptive beam-forming problem. The performance of the network was compared with other well-known complex-valued RBF networks available in literature, viz., split-complex CRBF, CMRAN and the CELM. The results indicate that the proposed fully complex-valued network has better convergence, approximation and classification ability.

This paper presents a reversible or lossless watermarking algorithm for images without using a location map in most cases. This algorithm employs prediction errors to embed data into an image. A sorting technique is used to record the prediction errors based on magnitude of its local variance. Using sorted prediction errors and, if needed, though rarely, a reduced size location map allows us to embed more data into the image with less distortion. The performance of the proposed reversible watermarking scheme is evaluated using different images and compared with four methods: those of Kamstra and Heijmans, Thodi and Rodriguez, and Lee et al. The results clearly indicate that the proposed scheme can embed more data with less distortion.

In a fully complex-valued feed-forward network, the convergence of the complex-valued back-propagation learning algorithm depends on the choice of the activation function, minimization criterion, initial weights and the learning rate. The minimization criteria used in the existing learning algorithms do not approximate the phase well in complex-valued function approximation problems. This aspect is very important in telecommunication and medical imaging applications. In this paper, we propose an improved complex-valued back propagation algorithm using an exponential activation function and a logarithmic minimization criterion, which approximates both the magnitude and phase well. Performance of the proposed scheme is evaluated using the complex XOR problem and a synthetic complex-valued function approximation problem. Also, a comparative analysis on the convergence of the existing fully complex and split complex networks is presented.

In this paper, we present a neural adaptive control scheme for active vibration suppression of a composite aircraft fin tip. The mathematical model of a composite aircraft fin tip is derived using the finite element approach. The finite element model is updated experimentally to reflect the natural frequencies and mode shapes very accurately. Piezo-electric actuators and sensors are placed at optimal locations such that the vibration suppression is a maximum. Model-reference direct adaptive neural network control scheme is proposed to force the vibration level within the minimum acceptable limit. In this scheme, Gaussian neural network with linear filters is used to approximate the inverse dynamics of the system and the parameters of the neural controller are estimated using Lyapunov based update law. In order to reduce the computational burden, which is critical for real-time applications, the number of hidden neurons is also estimated in the proposed scheme. The global asymptotic stability of the overall system is ensured using the principles of Lyapunov approach. Simulation studies are carried-out using sinusoidal force functions of varying frequency. Experimental results show that the proposed neural adaptive control scheme is capable of providing significant vibration suppression in the multiple bending modes of interest. The performance of the proposed scheme is better than the H(infinity) control scheme.

Visual tracking has been a challenging problem in computer vision over the decades. The applications of visual tracking are far-reaching, ranging from surveillance and monitoring to smart rooms. In this paper we present a novel object tracker based on fast learning radial basis function (RBF) networks. Here, the object and background pixel-based color features are used to develop object/non-object RBF classifiers. The posterior probability information of these classifiers are used for developing an efficient object model for tracking in the subsequent frames. The performance of the proposed tracker is tested with many video sequences of real-life complexity and compared against the color-based mean-shift tracker. The proposed tracker is illustrated to be suitable for real-time robust object tracking due to its low computational complexity.

Magnetic resonance imaging of dynamic events such as cognitive tasks in the brain, requires high spatial and temporal resolution. In order to increase the resolution in both domains simultaneously, parallel imaging schemes have been in existence, where multiple receiver coils are used, each of which needs to acquire only a fraction of the total available signal. In our approach, we regularly undersample dersample the signal at each of the receiver coils and the resulting aliased coil images are combined (unaliased) using the neural network framework. Data acquisition follows a variable-density sampling scheme, where lower frequencies are densely sampled, and the remaining signal is sparsely sampled. The low resolution images obtained using the densely sampled low frequencies are used to train the neural network. Reconstruction of the image is carried out by feeding the high-resolution aliased images to the trained network. The proposed approach has been applied to phantom as well as real brain MRI data sets, and results have been compared with the standard existing parallel imaging techniques. The proposed approach is found to perform better than the standard existing techniques.

In this paper, we present a machine learning approach to measure the visual quality of JPEG-coded images. The features for predicting the perceived image quality are extracted by considering key human visual sensitivity factors such as edge amplitude, edge length, background activity and background luminance. Image quality estimation involves computation of functional relationship between HVS features and subjective test scores. The subjective test scores for modified images are obtained with-out referring to their original images (called 'no reference'). Here, the problem of quality estimation is transformed to a sparse data classification problem using a sparse extreme learning machine (S-ELM). The S-ELM classifier estimate the posterior probability of a given image. Here, the mean opinion score ('visual quality') of an image is derived using the predicted class number and their estimated posterior probability. The experimental results prove that the estimated visual quality emulate the mean opinion score very well. The experimental results are compared with the existing JPEG no-reference image quality index and full-reference structural similarity image quality index. The result clearly shows the machine learning approach outperform the existing algorithms in the literature

This work presents an efficient mapping scheme for the multilayer perceptron (MLP) network trained using back-propagation (BP) algorithm on network of workstations (NOWs). Hybrid partitioning (HP) scheme is used to partition the network and each partition is mapped on to processors in NOWs. We derive the processing time and memory space required to implement the parallel BP algorithm in NOWs. The performance parameters like speed-up and space reduction factor are evaluated for the HP scheme and it is compared with earlier work involving vertical partitioning (VP) scheme for mapping the MLP on NOWs. The performance of the HP scheme is evaluated by solving optical character recognition (OCR) problem in a network of ALPHA machines. The analytical and experimental performance shows that the proposed parallel algorithm has better speed-up, less communication time, and better space reduction factor than the earlier algorithm. This work also presents a simple and efficient static mapping scheme on heterogeneous system. Using divisible load scheduling theory, a closed-form expression for number of neurons assigned to each processor in the NOW is obtained. Analytical and experimental results for static mapping problem on NOWs are also presented.