Hao Tang

Hallucination Begins Where Saliency Drops

MoSA: Motion-Coherent Human Video Generation via Structure-Appearance Decoupling

FOLIAGE: a Latent World Model for Accretive Surface Growth

VaseVQA-3D: Benchmarking 3D VLMs on Ancient Greek Pottery

SAYNEXT: A Benchmark and Cognitively Inspired Framework for Next-Utterance Prediction with Multimodal LLMs

Reinforcement Learning Agents in Quantum Code Discovery with Argmax-Preserving Quantization

3D-R1: Enhancing Reasoning in 3D VLMs for Unified Scene Understanding

MoRL: Reinforced Reasoning for Unified Motion Understanding and Generation

SpikeStereoNet: A Brain-Inspired Framework for Stereo Depth Estimation from Spike Streams

Fidelity-Aware Data Composition for Robust Robot Generalization

HGWM: Hierarchical Graph-guided World Model for Zero-shot Object Navigation via Scene-Goal Graph Matching

AnyDepth: Depth Estimation Made Easy

UniVid: The Open-Source Unified Video Model

Sim2Real-HOI: Sim-to-Real HOI Video Generation via Decoupled Motion–Appearance Diffusion

SafeMo: Trustworthy Motion Generation

Graph-Based Multi-Modal Light-weight Network for Adaptive Edge MRI Tumor Segmentation

DGS: Robust and Diverse Watermarks for Diffusion Models

Auth-Prompt Bench: Towards Reliable and Stable Prompting in Text-to-Image Generation

RFMedSAM 2: Automatic Prompt Refinement for Enhanced Volumetric Medical Image Segmentation with SAM 2

BagelScore: Visual-Language Evaluation Made Easy

SpatialVLA-Mamba: Efficient State-Space Models with Self-Refinement for Spatially-Grounded Robotic Control

RobustMerge: Parameter-Efficient Model Merging for MLLMs with Direction Robustness

Boosting Adversarial Transferability with Spatial Adversarial Alignment

Enhancing Diffusion-based Unrestricted Adversarial Attacks via Adversary Preferences Alignment

MaskSAM: Towards Auto-prompt SAM with Mask Classification for Medical Image Segmentation

M$^2$M: LEARNING CONTROLLABLE MULTI OF EXPERTS AND MULTI-SCALE OPERATORS ARE THE PARTIAL DIFFERENTIAL EQUATIONS NEED

Fusion-$\mathcal{X}$: Advancing LLM Ability with Adaptive Heterogeneous Model Integration