CMU Builds Self-Improving AI Agents via Constitutional RL
Carnegie Mellon researchers introduce Constitutional RL, a framework enabling AI agents to self-improve while following …
📂 Research
Latest articles in Research
KAIST Achieves Breakthrough in Energy-Efficient AI Chips
South Korean researchers develop novel AI chip architecture that cuts energy consumption by up to 95% compared to conven…
Toyota Research Uses GenAI to Speed Materials Discovery
Toyota Research Institute deploys generative AI models to dramatically accelerate the discovery of new materials for veh…
Sony Research Unveils Multimodal AI for Creative Content
Sony Research introduces a new multimodal AI model designed to generate and transform creative content across music, ima…
Microsoft MoE Architecture Slashes Inference Costs 70%
Microsoft Research unveils a sparse Mixture-of-Experts architecture that reduces AI inference costs by 70% while maintai…
OpenAI Advances Superalignment Research for AI
OpenAI publishes new research on superalignment techniques aimed at keeping frontier language models safe and aligned wi…
Stanford HAI Finds GPT-5 Shows Emergent Math Reasoning
A new Stanford HAI study reveals GPT-5 demonstrates unprecedented emergent reasoning abilities on complex mathematical p…
VinAI Advances Efficient Vision Transformers
Vietnam's VinAI Research publishes cutting-edge work on making Vision Transformers faster and lighter for real-world dep…
Toyota Research Institute Unveils Neural Net Breakthrough
TRI announces a major leap in autonomous driving neural networks, claiming 40% improvement in real-time decision-making …
NTT Unveils Photonic AI Chip That Slashes Power by 100x
Japan's NTT develops world-first photonic AI processor using light instead of electricity, achieving 100x lower power co…
Sony Research Unveils Multimodal AI for Creative Production
Sony Research has developed a new multimodal AI model designed to streamline creative content production across music, f…
MIT Sparse Attention Cuts LLM Inference Costs by 60%
MIT researchers unveil a new sparse attention mechanism that dramatically reduces LLM inference costs while preserving m…