We propose a federated reinforcement learning framework for autonomous navigation. The system aggregates policies from multiple vehicles while adapting to terrain diversity, allowing personalized learning without data exchange and improving real-time responsiveness in uncertain or evolving driving conditions.

Amit Krishna, Jean Marc Duval, Priya Manohar Reddy, Maria Gabriella Romero, Taro Kenji Nakamoto

This study presents a cross-modal learning framework for building emotion-aware agents. It fuses multi-resolution audio-visual signals with contextual language features, enhancing human-machine interaction through emotion-adaptive dialogue generation in varied affective states.

Robert James Whitmore, Laura Isabelle Simmons, Benjamin Luke Foster, Olivia Margaret Henderson, Rachel Claire Thornton

We propose a quantum regularized RNN for financial forecasting. It integrates noise-aware embedding optimization to improve stability in volatile sequences, enabling high-fidelity predictions across complex financial indicators with minimal sensitivity to outliers and non-stationary patterns.

Liu Cheng Hao, Zhang Rui Tao, Xu Ling Zhen, Gao Wen Ping, Huang Zhao Ming

This paper presents a privacy-preserving biometric verification framework using federated Siamese networks. The model uses adaptive contrastive learning to match identity across modalities without sharing sensitive data, enabling secure, decentralized authentication with robust performance across domains.

Ravi Ganesh Desai, Elena Maria Navarro, Siddharth Ramanujam Kulkarni, Carlos Eduardo Mendes, Jean Philippe Dufour

We introduce a hierarchical graph neural network for predicting drug-target interactions. The model supports scalable training and explainable predictions by leveraging layered graph aggregation and biochemical embeddings, advancing biomedical research and early-stage pharmaceutical screening.

James David Carter, Emily Alexandra Benson, Thomas Richard Wallace, Laura Helen Dawson, Olivia Jane Mitchell

This research introduces a bidirectional domain alignment framework for object detection under varying weather. It aligns visual distributions using generative techniques, enabling robust perception in autonomous driving systems exposed to diverse and unpredictable environmental conditions.

Chen Hao Wen, Liu Zhen Qiang, Xu Jian Ming, Zhang Tian Liang, Huang Rui Shan

This work introduces a temporal knowledge graph model for clinical event prediction. By learning from path-based attention over entity timelines, it enhances the reasoning of cause-effect relationships, supporting time-sensitive decision-making in digital healthcare systems with evolving patient data.

Chen Min Hao, Liu Fang Zhen, Zhang Xiu Wei, Xu Bo Wen, Gao Li Sheng, Huang Tian Jun