Abstract Nanopore sequencing offers the ability for real-time analysis of long DNA sequences at a low cost, enabling new applications such as early detection of cancer. Due to the complex nature of nanopore measurements and the high cost of obtaining ground-truth datasets, in-silico nanopore simulators play an important role in this field. Existing simulators rely on hand-crafted rules and parameters and do not learn an internal representation that would allow for analyzing underlying biological factors of interest. In this work, we investigate and extend Variational Autoregressive DNA-conditioned Autoencoder (VADA), a purely data-driven method for simulating nanopore signals based on an autoregressive latent variable model. We show that VADA can effectively model DNA-conditioned probability distributions over nanopore current sequences to produce varying current observations. We show that improving the flexibility of the conditional prior with conditional Real NVP flow significantly improves the simulation quality of VADA, a model that we refer to as VADA Normalizing Flow. We empirically demonstrate that our model achieves competitive simulation performance on experimental nanopore data. Moreover, we show that our model learns an informative latent representation that is predictive of the DNA labels. We hypothesize that other biological factors of interest, beyond DNA labels, can potentially be extracted from such a learned latent representation.

Neural network-based wireless receivers, also known as neural receivers , have demonstrated superior performance over traditional receivers, but come with greater computational complexity. The need to use these networks on energy-conscious edge devices is increasing, necessitating energy-efficient and adaptive neural receivers to function well under varying channel conditions. Transitioning static neural receivers to dynamic models using the concepts of Dynamic Neural Networks (DyNN) could reduce runtime computational complexity and energy consumption. This could be achieved by adapting the network architecture during run-time and exploit the varying channel conditions in a mobile communication system to reduce the computational complexity. This work introduces MEAN, a novel hard-gated Mixture-of-Experts (MOE) based neural receiver architecture. The main idea behind MEAN is to use several smaller Signal-to-Noise-Ratio (SNR) expert networks are used during run-time to create a network that selects the correct expert for the current data input to reduce complexity. The paper consists of the following key contributions. (1) MEAN architecture based on [ 25 ]: A hard-gated MoE model that dynamically selects the most suitable expert for each input dynamically based on current channel conditions. (2) Comprehensive system-level performance analysis of MEAN across different code rates and modulation schemes, including validation of expert selection during inference. (3) Loss function optimization for the gating network to promote the activation of specific experts in designated noise regions. This enhancement reduces gating network complexity while improving expert selection accuracy over previous MEAN implementations. (4) Hardware analysis of a gate-level implementation of MEAN synthesized in 22nm FD-SOI technology. The design is generated using a custom High-Level Synthesis (HLS) framework, providing detailed power and area evaluations. The proposed MEAN architecture for a Single Input Multiple Output (SIMO) wireless system achieves a 37.11% reduction in total power consumption with only an 11.04% increase in area, while maintaining accuracy comparable to that of a static neural network.

The cellular Potts model (CPM) is a powerful computational method for simulating collective spatiotemporal dynamics of biological cells. To drive the dynamics, CPMs rely on physics-inspired Hamiltonians. However, as first principles remain elusive in biology, these Hamiltonians only approximate the full complexity of real multicellular systems. To address this limitation, we propose NeuralCPM, a more expressive cellular Potts model that can be trained directly on observational data. At the core of NeuralCPM lies the Neural Hamiltonian, a neural network architecture that respects universal symmetries in collective cellular dynamics. Moreover, this approach enables seamless integration of domain knowledge by combining known biological mechanisms and the expressive Neural Hamiltonian into a hybrid model. Our evaluation with synthetic and real-world multicellular systems demonstrates that NeuralCPM is able to model cellular dynamics that cannot be accounted for by traditional analytical Hamiltonians.

Abstract Low-resource fake news detection aims at discerning between true and false claims from low-resource languages with scarce benchmark datasets. In this resource-constrained scenario, fake news data collected from online hoax reporting system is inherently skewed because human fact checkers mainly sample claims that are more likely to be fake or false. Instead of training end-to-end classifier on the extremely imbalanced dataset, our study investigates a simple framework based on contrastive learning and stacking-based ensemble learning as an alternate fake news classification pipeline for Indonesian language. Our empirical result shows that by combining contrastive-based embedding model—Contrast-BERT and ensemble of multilayer perceptrons (MLPs) in inference stage, we improve the precision score in fake news classification up to 26.64%, while maintaining accuracy and recall scores of above 75%, given extreme class imbalance ratio 1:24. Contrast-BERT is also superior to its counterparts in unsupervised topic clustering and evidence retrieval by nearly twofold. Furthermore, we observe that contrastive-based model follows a similar performance trend in Indonesian clickbait benchmark dataset. Contrast-BERT is more accurate and precise at predicting samples than end-to-end BERT classifier by up to 47%, given training subset with extreme imbalance ratio $$\ge$$ ≥ 1:19.

Dynamic Neural Networks (DyNN) adapt their structure during runtime for improved performance and lower power consumption. DyNNs benefit neural network-based wireless receivers, or in short neural receivers, that adapt their performance under varying channel conditions. However, DyNN architectures are not yet explored sufficiently in the literature as this would require a framework for simultaneously evaluating system-level performance and accurate hardware Power-Performance-Area (PPA) under different dynamic scenarios. This paper presents two main contributions: (1) An automated framework that bridges a system-level performance evaluation model of a wireless system and a High-Level Synthesis (HLS) tool for agile Design-Space Exploration (DSE) of DyNNs. (2) A novel DyNN architecture for neural receivers in wireless systems that skips a variable number of layers according to varying channel conditions. Our proposed neural receiver architecture with layer-skipping for a Single Input Multi Output (SIMO) wireless system when implemented in 22 nm FD-SOI technology shows power consumption savings of up to 59.2% at the cost of 200% increase in area compared to a static network.

Dynamic Neural Networks (DyNN) adapt its network architecture at run-time compared to static neural networks. DyNNs benefit in wireless receivers based on neural networks (or neural receivers), that need to adapt their performance under varying channel conditions. Mixture-of-Experts (MoE) is one efficient way to realize a DyNN in neural receivers in which several smaller expert networks are dynamically combined in the run-time to create a dedicated network according to the channel requirements. This paper presents a novel hard-gated (also known as sparsely-gated) MoE-based neural receiver architecture called MEAN for Single Input Multiple Output (SIMO)-based wireless communication systems. Our proposed MEAN architecture for a SIMO wireless system shows a reduction of up to 50% in the number of active layers during runtime.

Lottery tickets (LTs) is able to discover accurate and sparse subnetworks that could be trained in isolation to match the performance of dense networks. Ensemble, in parallel, is one of the oldest time-proven tricks in machine learning to improve performance by combining the output of multiple independent models. However, the benefits of ensemble in the context of LTs will be diluted since ensemble does not directly lead to stronger sparse subnetworks, but leverages their predictions for a better decision. In this work, we first observe that directly averaging the weights of the adjacent learned subnetworks significantly boosts the performance of LTs. Encouraged by this observation, we further propose an alternative way to perform an "ensemble'' over the subnetworks identified by iterative magnitude pruning via a simple interpolating strategy. We call our method Lottery Pools. In contrast to the naive ensemble which brings no performance gains to each single subnetwork, Lottery Pools yields much stronger sparse subnetworks than the original LTs without requiring any extra training or inference cost. Across various modern architectures on CIFAR-10/100 and ImageNet, we show that our method achieves significant performance gains in both, in-distribution and out-of-distribution scenarios. Impressively, evaluated with VGG-16 and ResNet-18, the produced sparse subnetworks outperform the original LTs by up to 1.88% on CIFAR-100 and 2.36% on CIFAR-100-C; the resulting dense network surpasses the pre-trained dense-model up to 2.22% on CIFAR-100 and 2.38% on CIFAR-100-C. Our source code can be found at https://github.com/luuyin/Lottery-pools.

In this study, we analyze automatic evaluation metrics for Natural Language Generation (NLG), specifically task-agnostic metrics and human-aligned metrics.Task-agnostic metrics, such as Perplexity, BLEU, BERTScore, are cost-effective and highly adaptable to diverse NLG tasks, yet they have a weak correlation with human.Human-aligned metrics (CTC, CtrlEval, UniEval) improves correlation level by incorporating desirable human-like qualities as training objective.However, their effectiveness at discerning system-level performance and quality of system outputs remain unclear.We present metric preference checklist as a framework to assess the effectiveness of automatic metrics in three NLG tasks: Text Summarization, Dialogue Response Generation, and Controlled Generation.Our proposed framework provides access: (i) for verifying whether automatic metrics are faithful to human preference, regardless of their correlation level to human; and (ii) for inspecting the strengths and limitations of NLG systems via pairwise evaluation.We show that automatic metrics provide a better guidance than human on discriminating system-level performance in Text Summarization and Controlled Generation tasks.We also show that multi-aspect human-aligned metric (UniEval) is not necessarily dominant over single-aspect human-aligned metrics (CTC, CtrlEval) and task-agnostic metrics (BLEU, BERTScore), particularly in Controlled Generation tasks. 1

Sparse training has received an upsurging interest in machine learning due to its tantalizing saving potential for both the entire training process as well as the inference. Dynamic sparse training (DST) as a leading approach can train deep neural networks at high sparsity from scratch to match the performance of their dense counterparts. However, most if not all DST prior arts demonstrate their effectiveness on unstructured sparsity with highly irregular sparse patterns, which receives limited support in common hardware. This limitation hinders the usage of DST in practice. In this paper, we propose Channel-aware dynamic sparse (Chase), that for the first time seamlessly translates the promise of unstructured dynamic sparsity to GPU-friendly channel-level sparsity (not fine-grained N:M or group sparsity) during one end-to-end training process, without any ad-hoc operations. The resulting small sparse networks can be directly accelerated by commodity hardware, without using any particularly sparsity-aware hardware accelerators. This appealing outcome is partially motivated by a hidden phenomenon of dynamic sparsity: off-the-shelf unstructured DST implicitly involves biased parameter reallocation across channels, with a large fraction of channels (up to 60%) being sparser than others. By progressively identifying and removing these channels during training, our approach transfers unstructured sparsity to channel-wise sparsity. Our experimental results demonstrate that Chase achieves 1.7x inference throughput speedup on common GPU devices without compromising accuracy with ResNet-50 on ImageNet. We release our code in https://github.com/luuyin/chase.

Neural networks are emerging as a tool for scalable data-driven simulation of high-dimensional dynamical systems, especially in settings where numerical methods are infeasible or computationally expensive. Notably, it has been shown that incorporating domain symmetries in deterministic neural simulators can substantially improve their accuracy, sample efficiency, and parameter efficiency. However, to incorporate symmetries in probabilistic neural simulators that can simulate stochastic phenomena, we need a model that produces equivariant distributions over trajectories, rather than equivariant function approximations. In this paper, we propose Equivariant Probabilistic Neural Simulation (EPNS), a framework for autoregressive probabilistic modeling of equivariant distributions over system evolutions. We use EPNS to design models for a stochastic n-body system and stochastic cellular dynamics. Our results show that EPNS considerably outperforms existing neural network-based methods for probabilistic simulation. More specifically, we demonstrate that incorporating equivariance in EPNS improves simulation quality, data efficiency, rollout stability, and uncertainty quantification. We conclude that EPNS is a promising method for efficient and effective data-driven probabilistic simulation in a diverse range of domains.

Automatically discovering a process model from an event log is the prime problem in process mining. This task is so far approached as an unsupervised learning problem through graph synthesis algorithms. Algorithmic design decisions and heuristics allow for efficiently finding models in a reduced search space. However, design decisions and heuristics are derived from assumptions about how a given behavioral description — an event log — translates into a process model and were not learned from actual models which introduce biases in the solutions. In this paper, we explore the problem of supervised learning of a process discovery technique. We introduce a technique for training an ML-based model using graph convolutional neural networks, which translates a given input event log into a sound Petri net. We show that training this model on synthetically generated pairs of input logs and output models allows it to translate previously unseen synthetic and several real-life event logs into sound, arbitrarily structured models of comparable accuracy and simplicity as existing state of the art techniques in imperative mining. We analyze the limitations of the proposed technique and outline alleys for future work.

The definition of Linear Symmetry-Based Disentanglement (LSBD) formalizes the notion of linearly disentangled representations, but there is currently no metric to quantify LSBD. Such a metric is crucial to evaluate LSBD methods and to compare them to previous understandings of disentanglement. We propose D_LSBD, a mathematically sound metric to quantify LSBD, and provide a practical implementation for SO(2) groups. Furthermore, from this metric we derive LSBD-VAE, a semi-supervised method to learn LSBD representations. We demonstrate the utility of our metric by showing that (1) common VAE-based disentanglement methods don’t learn LSBD representations, (2) LSBD-VAE, as well as other recent methods, can learn LSBD representations needing only limited supervision on transformations, and (3) various desirable properties expressed by existing disentanglement metrics are also achieved by LSBD representations.

Recent works on sparse neural network training have shown that a compelling trade-off between performance and efficiency can be achieved. Existing sparse training methods usually strive to find the best sparse subnetwork possible in one single run, without involving any expensive dense or pre-training steps. For instance, dynamic sparse training (DST), as one of the most prominent directions, is capable of reaching a competitive performance of dense training by iteratively evolving the sparse topology during the course of training. In this paper, we argue that it is better to allocate the limited resources to create multiple low-loss sparse subnetworks and superpose them into a stronger one, instead of allocating all resources entirely to find an individual subnetwork. To achieve this, two desiderata are required: (1) efficiently producing many low-loss subnetworks, the so-called cheap tickets, within one training process limited to the standard training time used in dense training; (2) effectively superposing these cheap tickets into one stronger subnetwork without going over the constrained parameter budget. To corroborate our conjecture, we present a novel sparse training approach, termed \textbf{Sup-tickets}, which can satisfy the above two desiderata concurrently in a single sparse-to-sparse training process. Across various models on CIFAR-10/100 and ImageNet, we show that Sup-tickets integrates seamlessly with the existing sparse training methods and demonstrates consistent performance improvement.

Deep neural networks are vulnerable to adversarial examples that are crafted by imposing imperceptible changes to the inputs. However, these adversarial examples are most successful in white-box settings where the model and its parameters are available. Finding adversarial examples that are transferable to other models or developed in a black-box setting is significantly more difficult. In this article, we propose the Direction-aggregated adversarial attacks that deliver transferable adversarial examples. Our method utilizes the aggregated direction during the attack process for avoiding the generated adversarial examples overfitting to the white-box model. Extensive experiments on ImageNet show that our proposed method improves the transferability of adversarial examples significantly and outperforms state-of-the-art attacks, especially against adversarial trained models. The best averaged attack success rate of our proposed method reaches 94.6% against three adversarial trained models and 94.8% against five defense methods. It also reveals that current defense approaches do not prevent transferable adversarial attacks.

Adversarial training is an approach of increasing the robustness of models to adversarial attacks by including adversarial examples in the training set. One major challenge of producing adversarial examples is to contain sufficient perturbation in the example to flip the model’s output while not making severe changes in the example’s semantical content. Exuberant change in the semantical content could also change the true label of the example. Adding such examples to the training set results in adverse effects. In this paper, we present the Calibrated Adversarial Training, a method that reduces the adverse effects of semantic perturbations in adversarial training. The method produces pixel-level adaptations to the perturbations based on novel calibrated robust error. We provide theoretical analysis on the calibrated robust error and derive an upper bound for it. Our empirical results show a superior performance of the Calibrated Adversarial Training over a number of public datasets.

Multi-Agent Path Finding (MAPF) is a routing problem in which multiple agents need to each find a lowest-cost collection of routes in a graph that avoids collisions between agents. This problem occurs frequently in the domain of logistics, for example in the routing of trains in shunting yards, airplanes at airports, and picking robots in automated warehouses. A solution is presented for the MAPF problem in which agents operate on an arbitrary directed graph, rather than the commonly assumed grid world, which extends support to use cases where the environment cannot be easily modeled in a grid shape. Furthermore, constraints are introduced on the start and end times of the routing tasks, which is vital in MAPF problems that are part of larger logistics systems. A Reinforcement Learning-based (RL) approach is proposed to learn a local routing policy for an agent in a manner that relieves the need for manually designing heuristics. It relies on a Graph Convolutional Network to handle arbitrary graphs. Both single-agent and multi-agent RL approaches are presented, showing how a multi-agent setup can reduce training time by exploiting the similarities in agent properties and local graph topologies.

Assigning meaning to parts of image data is the goal of semantic image segmentation. Machine learning methods, specifically supervised learning is commonly used in a variety of tasks formulated as semantic segmentation. One of the major challenges in the supervised learning approaches is expressing and collecting the rich knowledge that experts have with respect to the meaning present in the image data. Towards this, typically a fixed set of labels is specified and experts are tasked with annotating the pixels, patches or segments in the images with the given labels. In general, however, the set of classes does not fully capture the rich semantic information present in the images. For example, in medical imaging such as histology images, the different parts of cells could be grouped and sub-grouped based on the expertise of the pathologist. To achieve such a precise semantic representation of the concepts in the image, we need access to the full depth of knowledge of the annotator. In this work, we develop a novel approach to collect segmentation annotations from experts based on psychometric testing. Our method consists of psychometric testing procedure, active query selection, query enhancement, and a deep metric learning model to achieve a patch-level image embedding that allows for semantic segmentation of images. We show the merits of our method with evaluation on the synthetically generated image, aerial image and histology image.

Automatically discovering a process model from an event log is the prime problem in process mining. This task is so far approached as an unsupervised learning problem through graph synthesis algorithms. Algorithmic design decisions and heuristics allow for efficiently finding models in a reduced search space. However, design decisions and heuristics are derived from assumptions about how a given behavioral description – an event log – translates into a process model and were not learned from actual models which introduce biases in the solutions. In this paper, we explore the problem of supervised learning of a process discovery technique d. We introduce a technique for training an ML-based model d using graph convolutional neural networks; d translates a given input event log into a sound Petri net. We show that training d on synthetically generated pairs of input logs and output models allows d to translate previously unseen synthetic and several real-life event logs into sound, arbitrarily structured models of comparable accuracy and simplicity as existing state of the art techniques in imperative mining. We analyze the limitations of the proposed technique and outline alleys for future work.

Abstract Recurrent neural networks (RNNs) have achieved state-of-the-art performances on various applications. However, RNNs are prone to be memory-bandwidth limited in practical applications and need both long periods of training and inference time. The aforementioned problems are at odds with training and deploying RNNs on resource-limited devices where the memory and floating-point operations (FLOPs) budget are strictly constrained. To address this problem, conventional model compression techniques usually focus on reducing inference costs, operating on a costly pre-trained model. Recently, dynamic sparse training has been proposed to accelerate the training process by directly training sparse neural networks from scratch. However, previous sparse training techniques are mainly designed for convolutional neural networks and multi-layer perceptron. In this paper, we introduce a method to train intrinsically sparse RNN models with a fixed number of parameters and floating-point operations (FLOPs) during training. We demonstrate state-of-the-art sparse performance with long short-term memory and recurrent highway networks on widely used tasks, language modeling, and text classification. We simply use the results to advocate that, contrary to the general belief that training a sparse neural network from scratch leads to worse performance than dense networks, sparse training with adaptive connectivity can usually achieve better performance than dense models for RNNs.

Shortness of breath, or dyspnea is a condition of the cardio-pulmonary system that may be caused by, for example, a heart or lung disease, or physical load. In this paper, we explore techniques of detecting mild dyspnea directly from conversational speech, for example, in a telehealth application. We demonstrate with a collection of speech recordings before and after a light physical exercise that a siamese neural network, when presented examples of the two conditions, can detect the difference between two speech signals. This shows that this signal can be detected using data-pairs, removing the need for ratings of severity or the distinction of separate classes.

We examine deep neural network (DNN) performance and behavior using contrasting explanations generated from a semantically relevant latent space. We develop a semantically relevant latent space by training a variational autoencoder (VAE) augmented by a metric learning loss on the latent space. The properties of the VAE provide for a smooth latent space supported by a simple density and the metric learning term organizes the space in a semantically relevant way with respect to the target classes. In this space we can both linearly separate the classes and generate meaningful interpolation of contrasting data points across decision boundaries. This allows us to examine the DNN model beyond its performance on a test set for potential biases and its sensitivity to perturbations of individual factors disentangled in the latent space.

A standard Variational Autoencoder, with a Euclidean latent space, is structurally incapable of capturing topological properties of certain datasets. To remove topological obstructions, we introduce Diffusion Variational Autoencoders (DeltaVAE) with arbitrary (closed) manifolds as a latent space. A Diffusion Variational Autoencoder uses transition kernels of Brownian motion on the manifold. In particular, it uses properties of the Brownian motion to implement the reparametrization trick and fast approximations to the KL divergence. We show that the DeltaVAE is indeed capable of capturing topological properties for datasets with a known underlying latent structure derived from generative processes such as rotations and translations.

Facial micro-expressions are subtle and involuntary expressions that can reveal concealed emotions. Micro-expressions are an invaluable source of information in application domains such as lie detection, mental health, sentiment analysis and more. One of the biggest challenges in this field of research is the small amount of available spontaneous micro-expression data. However, spontaneous data collection is burdened by time-consuming and expensive annotation. Hence, methods are needed which can reduce the amount of data that annotators have to review. This paper presents a novel micro-expression spotting method using a recurrent neural network (RNN) on optical flow features. We extract Histogram of Oriented Optical Flow (HOOF) features to encode the temporal changes in selected face regions. Finally, the RNN spots short intervals which are likely to contain occurrences of relevant facial micro-movements. The proposed method is evaluated on the SAMM database. Any chance of subject bias is eliminated by training the RNN using Leave-One-Subject-Out cross-validation. Comparing the spotted intervals with the labeled data shows that the method produced 1569 false positives while obtaining a recall of 0.4654. The initial results show that the proposed method would reduce the video length by a factor of 3.5, while still retaining almost half of the relevant micro-movements. Lastly, as the model gets more data, it becomes better at detecting intervals, which the proposed method suitable for supporting the annotation process.

Unsupervised object discovery in images involves uncovering recurring patterns that define objects and discriminates them against the background. This is more challenging than image clustering as the size and the location of the objects are not known: this adds additional degrees of freedom and increases the problem complexity. In this work, we propose StampNet, a novel autoencoding neural network that localizes shapes (objects) over a simple background in images and categorizes them simultaneously. StampNet consists of a discrete latent space that is used to categorize objects and to determine the location of the objects. The object categories are formed during the training, resulting in the discovery of a fixed set of objects. We present a set of experiments that demonstrate that StampNet is able to localize and cluster multiple overlapping shapes with varying complexity including the digits from the MNIST dataset. We also present an application of StampNet in the localization of pedestrians in overhead depth-maps.

We present a method for detection of surface defects in images of 3D-printed products that enables automated visual quality control. The data characterising this problem is typically high-dimensional (high-resolution images), imbalanced (defects are relatively rare), and has few labelled examples. We approach these challenges by formulating the problem as probabilistic anomaly detection, where we use Variational Autoencoders (VAE) to estimate the probability density of non-faulty products. We train the VAE in an unsupervised manner on images of non-faulty products only. A successful model will then assign high likelihood to unseen images of non-faulty products, and lower likelihood to images displaying defects. We test this method on anomaly detection scenarios using the MNIST dataset, as well as on images of 3D-printed products. The demonstrated performance is related to the capability of the model to closely estimate the density distribution of the non-faulty (expected) data. For both datasets we present empirical results that the likelihood estimated with a convolutional VAE can separate the normal and anomalous data. Moreover we show how the reconstruction capabilities of VAEs are highly informative for human observers towards localising potential anomalies, which can aid the quality control process.

The integration of machine learning and inter- vehicle communications enables various active safety measures in internet-of-vehicles. Specifically, the environmental perception is processed by the deep learning module from vehicular sensor data, and the extended perception range is achieved by exchanging traffic-related information through inter-vehicle communications. Under such condition, the intelligent vehicles can not only percept the surrounding environment from self-collected sensor data, but also expand their perception range through the information sharing mechanism of Vehicular Ad-hoc Network (VANET). However, the dynamic urban environment in VANET leads to a number of issues, such as the effect of packet loss on the real-time perception accuracy of the received sensor data. In this work, we propose a point cloud object detection module via an end-to-end deep learning system and enable wireless communications between vehicles to enhance driving safety and facilitate real-time 3D mapping construction. Besides, we build a semi- realistic traffic scenario based on the Mong Kok district in Hong Kong to analyze the network performance of data dissemination under the dynamic environment. Finally, we evaluate the impact of data loss on the deep-learning-based object detection performance. Our results indicate that data loss beyond 50% (which is a common scene based on our simulation) can lead to a rapid decline of the object detection accuracy.

The Train Unit Shunting Problem (TUSP) is a difficult sequential decision making problem faced by Dutch Railways (NS). Current heuristic solutions under study at NS fall short in accounting for uncertainty during plan execution and do not efficiently support replanning. Furthermore, the resulting plans lack consistency. We approach the TUSP by formulating it as a Markov Decision Process and develop an image-like state space representation that allows us to develop a Deep Reinforcement Learning (DRL) solution. The Deep Q-Network efficiently reduces the state space and develops an on-line strategy for the TUSP capable of dealing with uncertainty and delivering significantly more consistent solutions compared to approaches currently being developed by NS.

Overhead depth map measurements capture sufficient amount of information to enable human experts to track pedestrians accurately. However, fully automating this process using image analysis algorithms can be challenging. Even though hand-crafted image analysis algorithms are successful in many common cases, they fail frequently when there are complex interactions of multiple objects in the image. Many of the assumptions underpinning the hand-crafted solutions do not hold in these cases and the multitude of exceptions are hard to model precisely. Deep Learning (DL) algorithms, on the other hand, do not require hand crafted solutions and are the current state-of-the-art in object localization in images. However, they require exceeding amount of annotations to produce successful models. In the case of object localization, these annotations are difficult and time consuming to produce. In this work we present an approach for developing pedestrian localization models using DL algorithms with efficient weak supervision from an expert. We circumvent the need for annotation of large corpus of data by annotating only small amount of patches and relying on synthetic data augmentation as a vehicle for injecting expert knowledge in the model training. This approach of weak supervision through expert selection of representative patches, suitable transformations and synthetic data augmentations enables us to successfully develop DL models for pedestrian localization efficiently.

Monitoring and controlling the user's Quality of Experience (QoE) in modern video services is a challenging proposition, mainly due to the limitations of current video quality assessment algorithms. While subjective QoE methods would better reflect the nature of human perception, these are not suitable in real-time automation cases. On the other hand, the existing objective algorithms are either too complex or too inaccurate, particularly in the context of lightweight devices such as camera sensors or smart phones. This paper introduces a novel objective QoE algorithm, Instantaneous Video Quality Assessment (IVQA), that is comparably as accurate as the most heavyweight algorithm available in the literature but can also be run in real-time. This approach is tested against a selection of ten objective metrics and benchmarked with a subjective user dataset.

With the widespread adoption of mobile Internet, the process of streaming video has become varied and complex. A diversity of factors affect the way we perceive quality in video streaming (also known as 'quality of experience', or QoE), involving far more than the individual video and network characteristics. Quality is affected by the overall delivery context, terminal specifications but also human factors. It is thus very hard to control the streaming system as a whole, targeting QoE rather than the tuning of individual factors. To better understand the non-obvious relation between network parameters and the resulting video quality, herein we present an experimental assessment of a representative video streaming platform, Skype. We find that simple QoE-management heuristics are only effective in very specific cases (for instance in 'head & shoulder' video types), which suggests that a more human-centric QoE management will be required to further improve video delivery.

Skype nodes generate a substantial part of real-time bi-directional video traffic nowadays. Employing a range of adaptive mechanisms, the application configures video streaming to meet the requirements of the communication and constraints of the underlying network. While other related works focus on passive network monitoring of Skype data flows, this paper studies Skype as a point-to-point streaming engine of video, beyond standard video calling. The emphasis is on the objective video quality as perceived by viewers. We built a testbed to generate network perturbations and stream certain videos between Skype nodes. We examine how network impairments affect objective metrics (i.e. PSNR and SSIM index) of the video and Skype's ability to reconstruct the original sample. The results suggest that Skype is weak at delivering good quality for high motion videos and that it is slow at recovering after a long period of high packet loss.

Measuring and predicting the user’s Quality of Experience (QoE) of a multimedia stream is the first step towards improving and optimizing the provision of mobile streaming services. This enables us to better understand how Quality of Service (QoS) parameters affect service quality, as it is actually perceived by the end user. Over the last years this goal has been pursued by means of subjective tests and through the analysis of the \nuser’s feedback.\nExisting statistical techniques have lead to poor accuracy (order of 70%) and inability to evolve prediction models with the system’s dynamics. In this paper, we propose a novel approach for building accurate and adaptive QoE prediction models using Machine\nLearning classification algorithms, trained on subjective test data.\nThese models can be used for real-time prediction of QoE and can be efficiently integrated into online learning systems that can adapt the models according to changes in the environment.\nProviding high accuracy of above 90%, the classification algorithms become an indispensible component of a mobile multimedia QoE management system.