This study proposes a multiscale residual learning model for super-resolving satellite images. By integrating generative refinement layers, the system reconstructs high-frequency details and mitigates blur, improving geospatial image clarity for land mapping, environmental monitoring, and spatial analytics applications.

Sarah Elizabeth Grant, James William Brooks, Olivia Charlotte Adams, Thomas Gregory Powell, Emily Catherine Bennett

We present a reinforcement learning-based scheduler for edge-cloud platforms. The system optimizes task distribution and reduces latency by learning energy-efficient execution policies, dynamically adjusting to load fluctuations and bandwidth variability in real-time collaborative computing environments.

Ravi Shankar Menon, Maria Lucia Navarro, Aditya Mohan Iyer, Jean Claude Bernard, Fatima Sofia Al-Sheikh

This research proposes a multi-task learning model for clinical text mining. It uses joint embedding and hierarchical label attention to extract medical insights while offering interpretability, supporting tasks like diagnosis classification, treatment recommendation, and symptom detection across varied medical records.

Michael Jonathan Ellis, Olivia Anne Porter, Sarah Abigail Clarke, Nathan Daniel Harris, Rachel Louise Foster

This paper presents an encoder-decoder network for real-time semantic segmentation in autonomous systems. It uses channel-wise attention and adaptive downsampling to improve inference speed and segmentation quality, enabling reliable navigation in resource-limited environments such as drones and mobile robots.

Zhang Hao Lin, Liu Feng Qiang, Xu Ming Zhao, Chen Jian Rui, Gao Wen Liang, Huang Zhi Ping

We propose a transformer-based relational inference model for completing sparse knowledge graphs. The architecture uses entity context encoding and path aggregation to predict missing facts, enhancing robustness and scalability in low-resource domains like medical ontologies and legal information systems.

Thomas Edward Hamilton, Emily Sophie Martin, Benjamin Patrick Young,

This paper presents a multilingual summarization system for legal documents using transformers. The model integrates contextual entity disambiguation and multi-head attention to generate concise and semantically coherent summaries, ensuring clarity and compliance across policy, legal, and governmental communication scenarios.

Sarah Elizabeth Matthews, Thomas Gregory Sanders, Olivia Lauren Jennings, Emily Catherine Wallace, Michael Andrew Grant

This paper introduces a graph transformer model for long-term traffic prediction. It captures complex temporal dynamics and spatial dependencies in urban networks, improving forecasting accuracy for transportation planning, route optimization, and infrastructure load balancing under real-world variability.

Chen Liang Xu, Zhang Hui Wen, Liu Min Hao, Gao Ping Zhao, Xu Rong Lin, Wang Zhen Tao