Eye feature extraction from event-based data streams can be performed efficiently and with low energy consumption, offering great utility to real-world eye tracking pipelines. However, few eye feature extractors are designed to handle sudden changes in event density caused by the changes between gaze behaviors that vary in their kinematics, leading to degraded prediction performance. In this work, we address this problem by introducing the adaptive inference state space model (AISSM), a novel architecture for feature extraction that is capable of dynamically adjusting the relative weight placed on current versus recent information. This relative weighting is determined via estimates of the signal-to-noise ratio and event density produced by a complementary dynamic confidence network. Lastly, we craft and evaluate a novel learning technique that improves training efficiency. Experimental results demonstrate that the AISSM system outperforms state-of-the-art models for event-based eye feature extraction.

Gastrointestinal cancer is one of the deadliest diseases all over the world. In medical imaging technology, cancer diagnosis has been evolving, especially endoscopes and currently applying artificial intelligence and deep learning on improving endoscopist’s imaging abilities. By using acquisition of images of tissues and organs, physicians are able to diagnose cancer in gastrointestinal track, besides of some limitations as depending on quality of endoscopy, experience of physicians or time consuming. With the development of technology, applications of deep learning are more effective in improving imaging techniques and diagnosis. Due to the state-of-the-art of deep learning, we applied VGG16, DenseNet201, HarDNet MSEG and HarDNet MSEG with attention to obtain some better results. Our results provide the comparison with mDice, mIoU and inference time on some of our trained models on the CVC-Clinic DB and Kvasir-SEG dataset. The best obtained mDice and mIoU are over 0.9 and 0.8, respectively.

Existing methods for merging experts during model training and fine-tuning predominantly rely on Euclidean geometry, which assumes a flat parameter space. This assumption can limit the model's generalization ability, especially during the pre-training phase, where the parameter manifold might exhibit more complex curvature. Curvature-aware merging methods typically require additional information and computational resources to approximate the Fisher Information Matrix, adding memory overhead. In this paper, we introduce CAMEx (Curvature-Aware Merging of Experts), a novel expert merging protocol that incorporates natural gradients to account for the non-Euclidean curvature of the parameter manifold. By leveraging natural gradients, CAMEx adapts more effectively to the structure of the parameter space, improving alignment between model updates and the manifold's geometry. This approach enhances both pre-training and fine-tuning, resulting in better optimization trajectories and improved generalization without the substantial memory overhead typically associated with curvature-aware methods. Our contributions are threefold: (1) CAMEx significantly outperforms traditional Euclidean-based expert merging techniques across various natural language processing tasks, leading to enhanced performance during pre-training and fine-tuning; (2) we introduce a dynamic merging architecture that optimizes resource utilization, achieving high performance while reducing computational costs, facilitating efficient scaling of large language models; and (3) we provide both theoretical and empirical evidence to demonstrate the efficiency of our proposed method. The code is publicly available at: https://github.com/kpup1710/CAMEx.

Although research has produced promising results demonstrating the utility of active inference (AIF) in Markov decision processes (MDPs), there is relatively less work that builds AIF models in the context of environments and problems that take the form of partially observable Markov decision processes (POMDPs). In POMDP scenarios, the agent must infer the unobserved environmental state from raw sensory observations, e.g., pixels in an image. Additionally, less work exists in examining the most difficult form of POMDP-centered control: continuous action space POMDPs under sparse reward signals. In this work, we address issues facing the AIF modeling paradigm by introducing novel prior preference learning techniques and self-revision schedules to help the agent excel in sparse-reward, continuous action, goal-based robotic control POMDP environments. Empirically, we show that our agents offer improved performance over state-of-the-art models in terms of cumulative rewards, relative stability, and success rate.

Music Emotion Recognition has attracted a lot of academic research work in recent years because it has a wide range of applications, including song recommendation and music visualization. As music is a way for humans to express emotion, there is a need for a machine to automatically infer the perceived emotion of pieces of music. In this paper, we compare the accuracy difference between music emotion recognition models given music pieces as a whole versus music pieces separated by instruments. To compare the models' emotion predictions, which are distributions over valence and arousal values, we provide a metric that compares two distribution curves. Using this metric, we provide empirical evidence that training Random Forest and Convolution Recurrent Neural Network with mixed instrumental music data conveys a better understanding of emotion than training the same models with music that are separated into each instrumental source.

This paper describes a deterministic approach to building a fixed-strategy gin rummy player. In the paper, we develop and evaluate both heuristic and neural network models for informing draw, discard, and knock decisions in the game. In this empirical study, we test performance of the models through competitive game play, show which best inform strategy, and demonstrate statistical significance of the improvement over a simple strategy. Through this empirical study, we indicate features that we expect to be helpful in future improvements to Gin Rummy play.

People detection in top-view, fish-eye images is challenging as people in fish-eye images often appear in arbitrary directions and are distorted differently. Due to this unique radial geometry, axis-aligned people detectors often work poorly on fish-eye frames. Recent works account for this variability by modifying existing anchor-based detectors or relying on complex pre/post-processing. Anchor-based methods spread a set of pre-defined bounding boxes on the input image, most of which are invalid. In addition to being inefficient, this approach could lead to a significant imbalance between the positive and negative anchor boxes. In this work, we propose ARPD, a single-stage anchor-free fully convolutional network to detect arbitrarily rotated people in fish-eye images. Our network uses keypoint estimation to find the center point of each object and regress the object’s other properties directly. To capture the various orientation of people in fish-eye cameras, in addition to the center and size, ARPD also predicts the angle of each bounding box. We also propose a periodic loss function that accounts for angle periodicity and relieves the difficulty of learning small-angle oscillations. Experimental results show that our method competes favorably with state-of-the-art algorithms while running significantly faster.

We first design an $\mathcal{O}(n^2)$ solution for finding a maximum induced matching in permutation graphs given their permutation models, based on a dynamic programming algorithm with the aid of the sweep line technique. With the support of the disjoint-set data structure, we improve the complexity to $\mathcal{O}(m + n)$. Consequently, we extend this result to give an $\mathcal{O}(m + n)$ algorithm for the same problem in trapezoid graphs. By combining our algorithms with the current best graph identification algorithms, we can solve the MIM problem in permutation and trapezoid graphs in linear and $\mathcal{O}(n^2)$ time, respectively. Our results are far better than the best known $\mathcal{O}(mn)$ algorithm for the maximum induced matching problem in both graph classes, which was proposed by Habib et al.

In this paper, we focus on broadening the monitoring area of a mobile air quality monitoring system, in which the sensors mounted on buses. In particular, we investigate the optimal buses to place the sensors and the optimal monitoring timings to maximize the number of monitored critical regions. We mathematically formulate the targeted problem. Then, we leverage the greedy approach to propose a polynomial-time$\frac{e-1}{2e-1}$approximation algorithm. We use the data of real bus routes in Hanoi, Vietnam, for the experimentation and show that the proposed algorithm guarantees an average performance ratio of 63.87%.

We introduce BAFT, a fast binary and quasi affine invariant local image feature. It combines the affine invariance of Harris Affine feature descriptors with the speed of binary descriptors such as BRISK and ORB. BAFT derives its speed and precision from sampling local image patches in a pattern that depends on the second moment matrix of the same image patch. This approach results in a fast but discriminative descriptor, especially for image pairs with large perspective changes. Our evaluation on 40 different image pairs shows that BAFT increases the area under the precision/recall curve (AUC) compared to traditional descriptors for the majority of image pairs. In addition we show that this improvement comes with a very low performance penalty compared to the similar ORB descriptor. The BAFT source code is available for download.

This paper presents the techniques employed in our team's submissions to the 2016 Emotion Recognition in the Wild contest, for the sub-challenge of group-level emotion recognition. The objective of this sub-challenge is to estimate the happiness intensity of groups of people in consumer photos. We follow a predominately bottom-up approach, in which the individual happiness level of each face is estimated separately. The proposed technique is based on geometric features derived from 49 facial points. These features are used to train a model on a subset of the HAPPEI dataset, balanced across expression and headpose, using Partial Least Squares regression. The trained model exhibits competitive performance for a range of non-frontal poses, while at the same time offering a semantic interpretation of the facial distances that may contribute positively or negatively to group-level happiness. Various techniques are explored in combining these estimations in order to perform group-level prediction, including the distribution of expressions, significance of a face relative to the whole group, and mean estimation. Our best submission achieves an RMSE of 0.8316 on the competition test set, which compares favorably to the RMSE of 1.30 of the baseline.