Yuxiang Liu

Yuxiang Liu

Tianjin University

Hi, I’m Yuxiang Liu (刘宇翔). Welcome to my homepage!

I am currently an undergraduate student at the College of Intelligence and Computing, Tianjin University (TJU).

My research interests include Computer Vision, 3D Reconstruction, and Embodied AI. I am actively involved in research under the supervision of Prof. Kun Li.

I have demonstrated strong capabilities in academic competitions and research. Notably, my team won first place in the Skeleton Tracking Challenge at CVPR 2025 Global 3D Human Poses (G3P) Workshop. I have also been awarded the National Scholarship for two consecutive years.

My Email: lyx1021@tju.edu.cn

🔥 News

  • 2026.08:  🚀 Released TurboT2VA, our framework for fast joint text-to-video-audio generation, achieving 54.67× generator-only speedup at high resolution on a single NVIDIA H20.
  • 2026.05:  🎉 Paper accepted by ICML 2026: EgoTSR — Evolving Ego-Centric Task-Oriented Spatiotemporal Reasoning via Curriculum Learning.
  • 2025.12:  ⭐ Awarded the National Scholarship for the academic year 2024-2025.
  • 2025.06:  🏆 Won first place in the Skeleton Tracking Challenge at the CVPR 2025 G3P Workshop.
  • 2024.12:  ⭐ Awarded the National Scholarship for the academic year 2023-2024.

📝 Publications

  • TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation
    Xiaoda Yang*, Yuxiang Liu*, Kaiwen Zheng, Yuan Liu, Yibo Lai, Shengpeng Ji, Kai Jiang, Jianfei Chen, Shan Yang, Sen Liang, Xiaobin Hu, Shuicheng Yan, Jintao Zhang†, Jun Zhu†, Zhou Zhao† (* equal contribution, † corresponding authors)
    arXiv preprint, 2026  [Paper | Code & Demos]
    • Proposed TurboT2VA, a score-regularized consistency distillation framework that accelerates a 19B-parameter joint video-audio model while preserving quality, diversity, and synchronization.
    • Distilled a 40-step teacher into a 4-step student with progressive discrete consistency warm-up, continuous consistency refinement, and joint consistency–distribution matching, achieving 20.1× generator speedup at 512×768 resolution.
    • Combined W8A8 quantization, fused operators, and modality-aware sparse attention to achieve 54.67× generator-only speedup at 1024×1792 resolution on a single NVIDIA H20 (318.74s → 5.83s).
  • From Perception to Planning: Evolving Ego-Centric Task-Oriented Spatiotemporal Reasoning via Curriculum Learning
    Xiaoda Yang*, Yuxiang Liu*, Shenzhou Gao, Can Wang, Jingyang Xue, Lixin Yang, Yao Mu, Tao Jin, Shuicheng Yan, Zhimeng Zhang, Zhou Zhao† (* equal contribution, † corresponding author)
    ICML 2026  [Paper | Code]
    • Proposed EgoTSR, a curriculum-based framework for learning task-oriented spatiotemporal reasoning that evolves from explicit spatial understanding to long-horizon planning.
    • Constructed EgoTSR-Data, a large-scale dataset comprising 46 million samples organized into three stages: CoT supervision, weakly supervised tagging, and long-horizon sequences.
    • Achieved 92.4% accuracy on long-horizon logical reasoning tasks, significantly outperforming existing open-source and closed-source state-of-the-art models.
  • Robust Camera Pose Estimation and 3D Human Reconstruction for Sports Events
    Jing Huang, Hanrong Zhuang, Lin Zhang, Yuxiang Liu, Kun Li
    Technical Report for FIFA Skeleton Light Challenge 2025  [Paper | Slides]
    • Proposed a method extending the RCR (Robust Crowd Reconstruction) framework to video inputs for sports events.
    • Designed a relative camera pose search algorithm with a fast line projector to achieve robustness and efficiency.
    • Refined 3D HVIP to ensure the consistency of human movement and extracted 3D skeletons from SMPL parameters.

🔬 Research Projects & Competitions

arXiv 2026
TurboT2VA: fast joint text-to-video-audio generation with 54.67× generator-only speedup

TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation

Paper | Code & Demos (arXiv 2026)

  • Accelerated a 19B-parameter joint video-audio model with progressive consistency distillation from 40 steps to 4 steps, preserving quality, diversity, and synchronization.
  • Achieved 20.1× generator speedup at 512×768 resolution through four-step distillation.
  • Combined distillation with W8A8 quantization, fused operators, and sparse attention for 54.67× generator-only speedup at 1024×1792 resolution on one NVIDIA H20.
ICML 2026
EgoTSR

EgoTSR: Evolving Ego-Centric Task-Oriented Spatiotemporal Reasoning via Curriculum Learning

arXiv 2604.10517 (ICML 2026)

  • Proposed a curriculum-based framework that evolves from explicit spatial understanding to internalized task-state assessment and long-horizon planning.
  • Constructed EgoTSR-Data with 46 million samples across three stages: CoT supervision, weakly supervised tagging, and long-horizon sequences.
  • Achieved 92.4% accuracy on long-horizon logical reasoning, outperforming existing SOTA models.
CVPR Winner
G3P Challenge

Champion of FIFA Innovation Challenge Skeleton Tracking Light

CVPR 2025 Global 3D Human Poses (G3P) Workshop

  • Achieved Rank 1 on the leaderboard. Invited to deliver the Winner Talk.
  • Developed a robust method for 3D skeleton tracking in complex unstructured scenarios.
Project Leader
Image Stitching

Unstructured Large Scene Light Field Image Stitching

National Innovation and Entrepreneurship Training Program (2024.08 - Present)

  • Advisor: Prof. Kun Li and Assistant Researcher Jian Ma.
  • Focusing on Trusted Projection and 3D Reconstruction for large-scale scenes.
  • Investigating algorithms for non-structured light field data processing.
Project
Multimodal Project

Multimodal Data Modeling & Knowledge Discovery

National Innovation and Entrepreneurship Training Program (2024.04 - 2025.04)

  • Core Member. Advisor: Prof. Yu Wang.
  • Status: Completed (One-year Project).
  • Implemented YOLO-based algorithms for license plate recognition under challenging conditions (high angle, blur, low light).
  • Achieved high accuracy in complex environments.

🏅 Honors and Awards

International & National

  • 2026 Paper Accepted, ICML 2026.
  • 2025.06 Champion, FIFA Skeleton Tracking Challenge (CVPR 2025 Workshop).
  • 2024–2025 academic year National Scholarship (Ministry of Education of China).
  • 2023–2024 academic year National Scholarship (Ministry of Education of China).

Provincial & Regional

  • 2025 First Prize, Lanqiao Cup National Software Talent Competition (C/C++, Tianjin Area).
  • 2024 First Prize, Tianjin Arts Performance (Orchestra).

🌟 Leadership & Activities

  • 2025.09 - Present: President, Student Union, School of Computer Science and Technology, TJU.
  • 2024.09 - Present: Vice Head, Peiyang Folk Orchestra.
  • 2024.09 - 2025.09: League Secretary, Top Talent Class (Bajian Class).

📖 Education

  • 2023.09 - Present, Undergraduate in Computer Science and Technology, Tianjin University (TJU).
    • Program: Top Talent Training Plan 2.0 (Bajian Class).
  • 2020.09 - 2023.07, High School, Shandong Qingdao No. 2 Middle School.
Total views:  |  Visitors:
Last updated: