Chuang Gan

Virtual Community: An Open World for Humans, Robots, and Society

BOAD: Discovering Hierarchical Software Engineering Agents via Bandit Optimization

PDTrack: Progressive Dense 3D Tracking

PhyScensis: Physics-Augmented LLM Agents for Complex Physical Scene Arrangement

EvoScale: Evolutionary Test-Time Scaling for Software Engineering

Steering LLM Thinking with Budget Guidance

Tailored Primitive Initialization is the Secret Key to Reinforcement Learning

Ella: Embodied Lifelong Learning Agents with Non-Parametric Memory

Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens

Deformable Linear Object Manipulations with Differentiable Physics

LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS

SafeDiffuser: Safe Planning with Diffusion Probabilistic Models

Scaling Autonomous Agents via Automatic Reward Modeling And Planning

ProsodyLM: Uncovering the Emerging Prosody Processing Capabilities in Speech Language Models

COMBO: Compositional World Models for Embodied Multi-Agent Cooperation

TopoGaussian: Inferring Internal Topology Structures from Visual Clues

Learning 3D Persistent Embodied World Models

MindJourney: Test-Time Scaling with World Models for Spatial Reasoning

RobotSmith: Generative Robotic Tool Design for Acquisition of Complex Manipulation Skills

DELTA: DENSE EFFICIENT LONG-RANGE 3D TRACKING FOR ANY VIDEO

ABNet: Adaptive explicit-Barrier Net for Safe and Scalable Robot Learning

AdaWorld: Learning Adaptable World Models with Latent Actions

Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search

CommVQ: Commutative Vector Quantization for KV Cache Compression

Orator: LLM-Guided Multi-Shot Speech Video Generation

Articulate Anything: Open-vocabulary 3D Articulated Object Generation

Virtual Community: A Generative Social World for Embodied AI

ABNet: Attention BarrierNet for Safe and Scalable Robot Learning

Learning 4D Embodied World Models

SnapMem: Snapshot-based 3D Scene Memory for Embodied Exploration and Reasoning