This study introduces a multi-stage reinforcement learning method to optimize adaptive energy control in smart grids. The model incorporates temporal prioritization and dynamic feedback mechanisms to enhance stability, scalability, and power distribution accuracy in complex, data-intensive infrastructure systems.

Amit Rajeev Bhatia, Jean Pierre Moreau, Priya Ramesh Nair, Carlos Fernando Gutierrez, Fatima Noor Al-Salim

This paper proposes a meta-learning framework that enables personalized diagnosis through dynamic context adaptation and episodic memory recall. It generalizes across heterogeneous patient profiles and delivers high accuracy even with limited clinical data for emerging or rare conditions.

Neha Pranav Iyer, Taro Masaki Nakamoto, Jean Claude Laurent, Pooja Sanjana Deshmukh, Elena Gabriela

This study introduces a generative model for visual question answering in low-resource languages. The framework fuses visual features with enriched semantic embeddings from multilingual sources, offering accurate comprehension and response generation without requiring extensive training corpora.

Sarah Elizabeth Turner, Emily Annette Brooks, Benjamin Thomas Clarke, Laura Katherine Simmons, Michael Andrew Foster

We propose a quantum transformer model for DNA sequence classification. Leveraging positionally-aware attention, it captures local and global dependencies efficiently, significantly improving accuracy and processing speed in genomic applications involving large-scale and high-dimensional biological sequences.

Liu Jin Rui, Xu Wei Han, Zhang Hui Long, Gao Ming Ze, Chen Bo Liang, Huang Xiang Feng

This paper presents a dual-stream CNN architecture for real-time semantic scene parsing. It integrates multi-level spatial alignment layers, enhancing boundary localization and object segmentation in autonomous driving tasks under variable lighting, scale, and surface conditions.

Robert Jacob Williams, Olivia Mae Thompson, Emily Claire Robertson, James Matthew Holland, Laura Grace Newman

We present a temporal graph convolutional model using residual attention for traffic prediction. The architecture captures spatiotemporal dependencies in urban networks, allowing for improved congestion forecasting and proactive routing across dynamic and high-density transportation systems.

Chen Wei Ming, Liu Fang Zhen, Zhang Rui Jun, Xu Long Tao, Gao Liang Hao, Huang Zhi Bo