Meta researchers open the LLM black box to repair flawed AI reasoning

The proof-of-concept could pave the way for a new class of AI debuggers, making language models more reliable for business-critical applications.
AZoBuild

AI Model Restores Lost Bridge Data to Improve Structural Safety

BiGRU, a deep learning model that enhances data recovery in structural health monitoring, ensuring the reliability of bridge ...
IEEE

Integrating Graph Neural Networks with Multi-Agent Deep Reinforcement Learning for Dynamic V2X Communication

Abstract: In the rapidly evolving field of Internet of Vehicles technology, Cellular Vehicle-to-Everything (C-V2X) communication has garnered considerable attention due to its superior performance in ...
Frontiers

Artificial intelligence and deep learning for neural data analysis

Artificial intelligence and deep learning have revolutionized the field of neural data analysis in recent years. The explosion of complex, high-dimensional ...
Frontiers

Advancing Computational Material Science and Mechanics through Integrated Deep Learning Models

The field of computational materials science has been profoundly transformed by integrating deep learning and other machine learning methodologies. These sophisticated data-driven approaches have ...

IBM's open source Granite 4.0 Nano AI models are small enough to run locally directly in your browser

IBM is entering a crowded and rapidly evolving market of small language models (SLMs), competing with offerings like Qwen3, ...

UAE researchers create AI systems that predict dementia and Alzheimer's

The second system, called MAGNET-AD, employs a graph neural network to detect Alzheimer’s disease before symptoms appear. It predicts both a patient’s cognitive performance score (PACC) and the time ...