Zhaoxiang Zhang

Full Professor@Institute of automation, Chinese academy of science, Chinese Academy of Sciences·中国·OpenReview

研究方向

Embodied Intelligence · Agent Learning · Autonomous driving · biological-inspired model · brain-inspired computing · deep neural networks · obejct detection · semantic segmentation · action recognition · person re-identification · scene analysis and understanding

Zhaoxiang Zhang

Unified Vision-Language-Action Model

IF-VidCap: Can Video Caption Models Follow Instructions?

YuE: Scaling Open Foundation Models for Long-Form Music Generation

Uniform Discrete Diffusion with Metric Path for Video Generation

DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving

ResGen: Residual Diffusion Model for LiDAR-based Point Cloud Generation

SpectraLLM: Uncovering the Ability of LLMs for Molecule Structure Elucidation from Multi-Spectra

Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Method

Grasp Any Region: Towards Precise, Contextual Pixel Understanding for Multimodal LLMs

GSPlane: Concise and Accurate Planar Reconstruction via Structured Representation

MLLM-CL: Continual Learning for Multimodal Large Language Models

FeatureBench: Benchmarking Agentic Coding for Complex Feature Development

HiPO: Hybrid Policy Optimization for Dynamic Reasoning in LLMs

OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs

From Genomic Whispers to Therapeutics: Multi-Resolution Transcriptome-Guided Diffusion Models for Drug Design and Screening

MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents

VGGT-X: When VGGT Meets Dense Novel View Synthesis

Scaling Autonomous Driving Safety with Synthetic Data

SIG-Chat: Spatial Intent-Guided Conversational Gesture Generation Involving How, When and Where

LVCap-Eval: Towards Holistic Long Video Caption Evaluation for Multimodal LLMs

IconBank: Mining Hard Samples via Visual Concepts for Data-Efficient GUI Grounding

KORGym: A Dynamic Game Platform for LLM Reasoning Evaluation

Enhancing End-to-End Autonomous Driving with Latent World Model

DriveDPO: Policy Learning via Safety DPO For End-to-End Autonomous Driving

TC-Light: Temporally Coherent Generative Rendering for Realistic World Transfer

CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes

McEval: Massively Multilingual Code Evaluation

FreeVS: Generative View Synthesis on Free Driving Trajectory

Reconstructive Visual Instruction Tuning

MTU-Bench: A Multi-granularity Tool-Use Benchmark for Large Language Models

OmniBench: Towards The Future of Universal Omni-Language Models

MIO: A Foundation Model on Multimodal Tokens

POC: Preventing the Over-Collapse of Classes for Class-Incremental Learning

Towards Flexible and Controllable Unknown Rejection

AutoGUI: Scaling GUI Grounding with Automatic Functionality Annotations from LLMs

HelloBench: Evaluating Long Text Generation Capabilities of Large Language Models

FlexDrive: Toward Trajectory Flexibility in Driving Scene Reconstruction and Rendering

UI-Pro: A Hidden Recipe for Building Vision-Language Models for GUI Grounding