Zhuotao Tian

Full Professor@Harbin Institute of Technology (Shenzhen)·中国·OpenReview

研究方向

Multi-modal perception · Large Language Models · Multi-modal understanding and analysis · Few-shot segmentation and Few-shot Learning · Semantic Segmentation · Computer Vision

Zhuotao Tian

Efficient Reasoning with Balanced Thinking

PointRePar : SpatioTemporal Point Relation Parsing for Robust Category-Unified 3D Tracking

SafeMVDrive: Multi-view Safety-Critical Driving Video Generation in the Real World Domain

Dynamic-dLLM: Dynamic Cache-Budget and Adaptive Parallel Decoding for Training-Free Acceleration of Diffusion LLM

FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token Merging

Plug-and-Play Fidelity Optimization for Diffusion Transformer Acceleration via Cumulative Error Minimization

Video-ToC: Video Tree-of-Cue Reasoning

Uni-DPO: A Unified Paradigm for Dynamic Preference Optimization of LLMs

GRASP-GS: Geometric Registration and Dual-Stag Saliency Pruning for Efficient 3D Gaussian Splatting

ES-GGT: Efficient Submap-based Visual Geometry Grounded Transformer with Spatial Memory Alignment

Multimodal Dataset Distillation via Phased Teacher Models

LongHorizonUI: A Unified Framework for Robust long-horizon Task Automation of GUI Agent

Memory Forcing: Spatio-Temporal Memory for Consistent Scene Generation on Minecraft

Consistency Beyond Contrast: Enhancing Object Detection Robustness through Scene-Augmented Feature Alignment

VisionFocus: Towards Efficient Hallucination Mitigation via Token-Aware Visual Enhancement

PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval

Amodal SAM: Open-World Amodal Segmentation

Concerto: Joint 2D-3D Self-Supervised Learning Emerges Spatial Representations

Context-Aware Hierarchical Learning: A Two-Step Paradigm towards Safer LLMs

Less Is More, but Where? Dynamic Token Compression via LLM-Guided Keyframe Prior

SR$^2$: BOOSTING 3D LARGE LANGUAGE MODEL WITH SPATIAL RELATION REASONING

Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs