Blog
No more endless PDFs. Discover the core value of the latest top-tier research in one article.
SAW-INT4: System-Aware 4-Bit KV-Cache Quantization for Real-World LLM Serving
PLaMo 2.1-VL Technical Report
The Flat Critical Branch Between Nariai and Bertotti-Robinson Geometries as a Solution of Cosmological Einstein-Maxwell Theory
Micro Language Models Enable Instant Responses
Architecting Early Fault Tolerant Neutral Atoms Systems with Quantum Advantage
PortraitDirector: A Hierarchical Disentanglement Framework for Controllable and Real-time Facial Reenactment
Synchronization in a dissipative quantum many-body system
From Craft to Kernel: A Governance-First Execution Architecture and Semantic ISA for Agentic Computers
AI scientists produce results without reasoning scientifically
HALO: Hybrid Auto-encoded Locomotion with Learned Latent Dynamics, Poincaré Maps, and Regions of Attraction
Oort Cloud Ecology - IV. Exchanging Asteroids
Image Generators are Generalist Vision Learners
Generalization at the Edge of Stability
LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model
Wan-Image: Pushing the Boundaries of Generative Visual Intelligence
TEMPO: Scaling Test-time Training for Large Reasoning Models
Near-Future Policy Optimization
FASTER: Value-Guided Sampling for Fast RL
PokeVLA: Empowering Pocket-Sized Vision-Language-Action Model with Comprehensive World Knowledge Guidance
JoyAI-RA 0.1: A Foundation Model for Robotic Autonomy
RoboWM-Bench: A Benchmark for Evaluating World Models in Robotic Manipulation
Exploring Spatial Intelligence from a Generative Perspective
CityRAG: Stepping Into a City via Spatially-Grounded Video Generation
Mask World Model: Predicting What Matters for Robust Robot Policy Learning
A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression
X-Cache: Cross-Chunk Block Caching for Few-Step Autoregressive World Models Inference
UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling
VLA Foundry: A Unified Framework for Training Vision-Language-Action Models
DR-Venus: Towards Frontier Edge-Scale Deep Research Agents with Only 10K Open Data
MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings