This research tackles Timescale Dependent Label Inconsistency in EEG emotion recognition by introducing two innovative regularization strategies. The methods enhance model performance and explainability, validated through experiments on popular datasets. Results show significant improvements over existing models, emphasizing a balance between interpretability and predictive accuracy.

This paper explores a new approach to stochastic optimization, combining non-smooth outer functions with smooth inner ones. Through the use of smoothing functions, the authors present a method that enhances convergence rates in various settings, demonstrating competitive results for specific problem classes in machine learning.

LiLM-RDB-SFC presents an innovative method for optimizing Service Function Chain (SFC) provisioning using a combination of Lightweight Language Models and Relational Databases in a Deep Reinforcement Learning framework. The approach enhances decision-making in Software-Defined Networking by improving accuracy and reducing processing time significantly.

This article explores the concept of class-difficulty separability in machine learning, highlighting how data difficulty varies across classes. The authors present a new method for data pruning that enhances model performance, particularly in complex domains like network intrusion detection and medical imaging, achieving state-of-the-art results.

This study evaluates the effectiveness of a Large Language Model (LLM)-based multi-agent system for therapy recommendations in chronic patients with multimorbidity. It compares the performance of multi-agent systems to single-agent approaches, revealing strengths and limitations in providing safe, comprehensive treatment advice.

HanjaBridge introduces a new technique for improving Korean language understanding in large language models by addressing semantic ambiguity. By integrating Hanja characters into pre-training, this method facilitates better contextual learning and enhances model performance, achieving significant improvements on benchmarks while ensuring practical efficiency.

Protein language models offer significant benefits in generating sequences, but they carry risks of harmful outcomes. This article presents a Knowledge-guided Preference Optimization framework to enhance the safety of protein generation while preserving functionality, addressing crucial biosafety and ethical challenges in biotechnology.

This article presents TableEG, a framework that uses large language models to generate realistic errors in tabular data. By enabling authentic error creation for better training of detection algorithms, it addresses the challenge of evaluating data cleaning techniques and enhances the understanding of data quality issues in various domains.

This study evaluates how well Large Language Models (LLMs) understand story-based analogies compared to humans. It investigates their ability to capture semantic similarities, explore reasoning through prompts, and assesses the effects of model size, revealing insights into LLMs' potential for human-like reasoning.

This article introduces the WomenSports dataset, aimed at improving sports action classification specifically for women. It features diverse sports activities and utilizes a convolutional neural network for deep feature extraction, achieving impressive classification accuracy. The dataset is publicly available, enhancing research opportunities in this area.

This article explores a new approach to human-object interaction detection by introducing a multi-scale wavelet attention backbone and ray-based encoder. These innovations aim to improve prediction accuracy and efficiency while addressing limitations in existing methods. Experimental results indicate promise with benchmark datasets.

This review explores the role of Physics-Informed Neural Networks in improving semiconductor film deposition processes. It highlights recent advancements, identifies research gaps, and suggests future directions for integrating machine learning with physical principles to enhance accuracy, efficiency, and scalability in semiconductor manufacturing.

Scaling reinforcement learning through distributed simulation can be challenging. This article introduces ClusterEnv, a flexible framework that enhances modularity by decoupling simulation from training. Featuring Adaptive Actor Policy Synchronization, it efficiently manages policy updates while integrating easily into existing RL systems, all with minimal changes.

This article examines the misalignment in reward functions used in machine learning, highlighting how blending instrumental and terminal goals can lead to poor outcomes. It presents a straightforward example illustrating this issue and discusses its implications for reinforcement learning in various environments.

This study addresses the unique challenges of adversarial attacks on tabular data, which differ from those in image and text domains. By using a Variational Autoencoder framework, the authors generate imperceptible adversarial examples while preserving statistical properties, achieving improved performance across multiple datasets and models.