Jiangmiao Pang

$\pi^3$: Permutation-Equivariant Visual Geometry Learning

RoboInter: A Holistic Intermediate Representation Suite Towards Robotic Manipulation

SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning

MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Ground Slow, Move Fast: A Dual-System Foundation Model for Generalizable Vision-Language Navigation

Vision-Language-Action Instruction Tuning: From Understanding to Manipulation

ARTDECO: Toward High-Fidelity On-the-Fly Reconstruction with Hierarchical Gaussian Structure and Feed-Forward Guidance

OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modeling

Spatially Guided Training for Vision-Language-Action Model

ChangingGrounding: 3D Visual Grounding in Changing Scenes

Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies

Vlaser: Vision-Language-Action Model with Synergistic Embodied Reasoning

DeepVerse: 4D Autoregressive Video Generation as a World Model

Expertise need not monopolize: Action-Specialized Mixture of Experts for Vision-Language-Action Learning

VLAC: A Generalist Action-Critic Model via Pair-wise Progress Understanding

MV-CoLight: Efficient Object Compositing with Consistent Lighting and Shadow Generation

MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning

Predictive Inverse Dynamics Models are Scalable Learners for Robotic Manipulation

MeshCoder: LLM-Powered Structured Mesh Code Generation from Point Clouds

EgoThinker: Unveiling Egocentric Reasoning with Spatio-Temporal CoT

Evolving Symbolic 3D Visual Grounder with Weakly Supervised Reflection

An Image is Worth $K$ Slots: Data-efficient Scaling of Self-supervised Visual Pre-training