Global Intelligence Database

Results for "Computer Science & AI"

Front Page Only

FOUND 88 MATCHESTIME: 398ms

#71Computer Science & AIFront Page27 April 2026

Optimising Multi-Agent Reinforcement Learning Through Conditional Policy Decomposition

Researchers propose CTMAPPO-Clip to prevent AI agents from memorising specific training scenarios. By using advantage clipping and Transformer architectures, the system aims to improve coordination stability within centralised execution environments.

By Bo, Wang, Han, Shi, Niu, Li, Zhang, Cui

#72Computer Science & AIFront Page25 April 2026

A New Computational Benchmark for Early Alzheimer's disease diagnosis

ANA-GNN uses adaptive graph modelling to identify early cognitive impairment with 85.23% accuracy. By prioritising specific biological regions like the hippocampus, it outperforms previous computational benchmarks.

By Sheng, Zhong, Zhang, Zhang, Gong, Lin, Chen

Page 8 of 9