This study presents a cross-platform reinforcement learning approach for smart traffic signal control. The system uses deep Q-learning and state representation compression to adapt signal strategies under variable demand, reducing congestion while maintaining stability across different urban environments.

Zhang Wei Ming, Li Jian Rong, Xu Qing Hao, Huang Chen Xi, Liu Rui Zhe

A quantum-regularized kernel learning model is developed for imbalanced datasets with noisy features. The method stabilizes margin optimization, improves minority class recognition, and ensures generalization in noisy settings using a regularized quantum distance metric integrated into the SVM formulation.

Wei Zhen Liang, Liu Feng Ming, Zhang Han Jie, Xu Dong Qiang, Gao Lin Yue, Yang Mei Xiu

We propose a meta-learned attention filtering module that enhances few-shot learning across visual domains. The mechanism identifies transferable features and suppresses irrelevant activations, improving classification performance with limited samples and supporting rapid adaptation across disjoint domain settings.

Emily Samantha Collins, Ryan Joseph Matthews, Benjamin Henry Clark, Olivia Rose Martin, Nathan James Grant

This paper presents a hierarchical reinforcement learning model with policy refinement for autonomous navigation. It enables agents to adapt to uncertain and dynamic environments using layered decision policies and feedback loops, improving path reliability and minimizing failure in obstacle-rich spaces.

Ravi Harish Kulkarni, Ivan Dmitri Petrov, Priya Malini Subramanian, Carlos Eduardo Moreno, Anna Isabelle Müller

This study introduces a deep transfer learning model that improves interpretability and adaptability in high-dimensional sensor environments. The framework enhances domain transferability and reduces calibration overhead, making it effective for predictive maintenance, operational forecasting, and adaptive decision-making in complex deployments.

Amit Rajan Kulkarni, Jean Michel Laurent, Fatima Noor Al-Mutlaq, Priya Devi Sharma, Carlos Manuel Ortiz

This paper presents a scalable quantum neural framework for real-time analysis of multichannel biomedical imaging. It combines quantum feature mapping and entanglement-inspired learning to recognize patterns in noisy environments, improving precision, interpretability, and computational efficiency under constrained clinical conditions.

Sarah Elizabeth Jenkins, Michael Thomas Grant, Olivia Katherine Wallace, Benjamin Matthew Scott, Laura Christine Adams

We introduce a distributed multi-agent deep learning system for energy-aware smart grid optimization. The architecture employs agent-level learning with cooperative reward shaping, ensuring energy efficiency, fault resilience, and distributed control across multiple generation and consumption nodes.

Michael Jonathan Lewis, Rachel Anna Thompson, Thomas Gregory Ellis, Laura Natalie Evans, Robert Edward Young