Exotic Particles Could Upend the Standard Model as AI Accelerates New Discoveries in Particle Physics
Physicists are harnessing AI to hunt for exotic particles beyond the Standard Model. From tetraquarks to magnetic monopo…
📂 Research
Latest articles in Research
Study Reveals: Humans Default to Trusting AI Opponents as More Rational
The first controlled laboratory study finds that in strategic games, humans tend to assume large language model opponent…
"Zealot" Simulated Attack Reveals AI Cyber Threats Far Exceed Expectations
Security researchers demonstrated AI-driven cyberattack capabilities in a simulated cloud environment through a proof-of…
Breakthroughs in Large Model Quantization Algorithms
As large language model parameters continue to scale up, advanced quantization algorithms have become a critical technic…
Xmemory Benchmark: Structured AI Memory Challenges the RAG Paradigm
The new Xmemory benchmark framework is the first to systematically compare structured AI memory against RAG and hybrid R…
Zero-Shot Transfer Breakthrough: World Model Driving Agents Generalize Across Cities
A new paper proposes Semantic Rollout and Town-Adversarial Regularization methods, achieving zero-shot cross-town transf…
MotuBrain: A World Action Model That Unifies Video and Action Modeling for Robots
A latest arXiv paper introduces MotuBrain, which jointly models video and actions based on the UniDiffuser framework, br…
A Novel Robot Localization Method Based on Hierarchical Scene Graph Matching
A latest study proposes a learning-based hierarchical scene graph matching method that achieves precise indoor robot loc…
Active Inference Framework Enables Autonomous Trajectory Planning for Drone Swarms
A new study proposes a hierarchical probabilistic reasoning framework based on Active Inference, transforming multi-UAV …
ExoActor: Driving Humanoid Robot Interaction Control Through Exocentric Video Generation
A latest arXiv paper proposes the ExoActor framework, which redefines exocentric video generation as a generalizable hum…
Survey: How Robots Learn Manipulation Skills from Human Videos
A latest arXiv survey systematically reviews research progress on robots learning manipulation skills from human video d…
"Connected Dependability Cage": A New Framework for Runtime Safety Monitoring in Autonomous Driving
A German research team has proposed the "Connected Dependability Cage" concept, which provides full-lifecycle safety ass…