Yuxiong Wang

Yuxiong WANG



Postdoctoral Fellow

Robotics Institute, School of Computer Science, Carnegie Mellon University

Email: yuxiongw[at]cs[dot]cmu[dot]edu

Google Scholar


Make things as simple as possible, but not simpler

-- Albert Einstein

Welcome to my homepage. I am a postdoctoral fellow in the Robotics Institute, School of Computer Science, at Carnegie Mellon University advised by Prof. Martial Hebert. I obtained my Ph.D. under the supervision of Prof. Martial Hebert in the Robotics Institue in May, 2018. I have also been closely working with Prof. Deva Ramanan and Prof. Ruslan Salakhutdinov. Starting in fall semester, 2019, I have been visting the Center for Data Science at New York University, working with Prof. Jean Ponce. I did a recent internship with Dr. Ross Girshick and Prof. Bharath Hariharan at Facebook AI Research (FAIR). Previously, I got my Master's degree at Tsinghua University advised by Prof. Yu-Jin Zhang, and received my B.E. at Beijing Institute of Technology, both with highest honor.

My research lies in computer vision, machine learning, and robotics, with a specific focus on meta-learning and few-shot learning.

Current Research Topics

  • Meta-learning and learning to learn
  • Few/low-shot recognition and detection, long-tail recognition
  • Generative modeling, predictive learning
  • Continual learning, transfer learning, domain adaptation
  • Large-scale unsuperivsed, discriminative learning
  • Human motion prediction for human-robot interaction

Dissertation

Selected Publications (by date / by topic)

Task-Oriented & Few-Shot Generative Modeling

    Progressive Knowledge Distillation for Few-Shot Learning

    Few-shot GAN

    Few-Shot Generative Adversarial Networks

    Chunyan Bai, Yan Xu, Boyu Liu, Ruslan Salakhutdinov, Martial Hebert, Yuxiong Wang

    Under review, 2020. [Preprint]


    Meta-Learning by Hallucinating Useful Examples

    Meta-Learning by Hallucinating Useful Examples

    Yuxiong Wang*, Yuki Uchiyama*, Martial Hebert, Karteek Alahari

    Under review, 2020. [Preprint]


    Image Deformation Meta-Networks for One-Shot Learning

    Image Deformation Meta-Networks for One-Shot Learning

    Zitian Chen, Yanwei Fu, Yuxiong Wang, Lin Ma, Wei Liu, Martial Hebert

    Oral Presentation, Best Paper Award Finalist, CVPR, 2019. [PDF][Talk]


    Embodied One-Shot Video Recognition: Learning from Actions of a Virtual Embodied Agent

    Embodied One-Shot Video Recognition: Learning from Actions of a Virtual Embodied Agent

    Yuqian Fu, Chengrong Wang, Yanwei Fu, Yuxiong Wang, Cong Bai, Xiangyang Xue, Yu-Gang

    Jiang, Lin Ma, Wei Liu, Martial Hebert

    ACM MM, 2019. [PDF]


    Task-Oriented Generative Modeling

    Low-Shot Learning from Imaginary Data

    Yuxiong Wang, Ross Girshick, Martial Hebert, Bharath Hariharan

    Spotlight Oral Presentation, CVPR, 2018. [PDF][Poster][Talk]

Streaming Image Understanding

    Towards Streaming Image Understanding

    Towards Streaming Image Understanding

    Mengtian Li, Yuxiong Wang, Deva Ramanan

    Under review, 2020. [Preprint]

3D Human Motion Prediction and Its Application in Human-Robot Interaction

    Stochastic Dual-Attention Long-Term Motion Prediction

    Learning to Predict Diverse Futures from a Single Past Motion Sequence

    Yuxiong Wang*, Liang-Yan Gui*, José M. F. Moura

    Under review, 2019 (* indicates equal contribution). [Preprint]


    Motion Prediction for Human-Robot Interaction

    Teaching Robots to Predict Human Motion

    Liang-Yan Gui, Kevin Zhang, Yuxiong Wang, Xiaodan Liang, José M. F. Moura, Manuela M. Veloso

    Oral Presentation, IROS, 2018. [PDF]


    Few-Shot Motion Prediction via Meta-Learning

    Few-Shot Human Motion Prediction via Meta-Learning

    Liang-Yan Gui, Yuxiong Wang, Deva Ramanan, José M. F. Moura

    ECCV, 2018. [PDF][Poster]


    Human-Like Motion Prediction

    Adversarial Geometry-Aware Human Motion Prediction

    Yuxiong Wang*, Liang-Yan Gui*, Xiaodan Liang, José M. F. Moura

    Oral Presentation, ECCV, 2018 (* indicates equal contribution). [PDF][Poster][Talk]

Compositional Learning

    Learning Generalizable Representations via Diverse Supervision

    Learning Generalizable Representations via Diverse Supervision

    Ziqi Pang, Zhiyuan Hu, Pavel Tokmakov, Yuxiong Wang, Martial Hebert

    Under review, 2020. [arXiv]


    Compositional Learning

    Learning Compositional Representations for Few-Shot Recognition

    Pavel Tokmakov, Yuxiong Wang, Martial Hebert

    ICCV, 2019. [PDF][Poster]

Knowledge Distillation and Learning to Learn in Model Space

    Meta-Learning to Detect Rare Objects

    Meta-Learning to Detect Rare Objects

    Yuxiong Wang, Deva Ramanan, Martial Hebert,

    ICCV, 2019. [PDF][Poster]


    Few-Shot Motion Prediction via Meta-Learning

    Few-Shot Human Motion Prediction via Meta-Learning

    Liang-Yan Gui, Yuxiong Wang, Deva Ramanan, José M. F. Moura

    ECCV, 2018. [PDF][Poster]


    Meta-Learning of Model Dynamics for Long-Tail Recognition

    Learning to Learn through Model Regression Networks

    Learning to Learn: Model Regression Networks for Easy Small Sample Learning

    Yuxiong Wang, Martial Hebert

    ECCV, 2016. [PDF][Poster]

Domain Adaptation

    Prototypical Adaptation for Few-Shot Classification

    Prototypical Adaptation for Few-Shot Classification

    Rajshekhar Das, Yuxiong Wang, José M. F. Moura

    Under review, 2020. [Preprint]


    Transfer Learning from Unsupervised Universal Sources

    Learning by Transferring from Unsupervised Universal Sources

    Yuxiong Wang, Martial Hebert

    Oral Presentation, AAAI, 2016. [PDF]

Rethinking Fine-Tuning via Developmental Learning

    Continual Learning and Fine-Tuning by Increasing Model Capacity

    Unsupervised Fine-Tuning

    Factorized Convolutional Networks: Unsupervised Fine-Tuning for Image Clustering

    Liang-Yan Gui, Liangke Gui, Yuxiong Wang, Louis-Philippe Morency, José M. F. Moura

    Oral Presentation, WACV, 2018. [PDF]

Unsupervised Meta-Learning

    Improving CNN Transferability through Large-Scale Unsupervised Meta-Training

    Learning from Small Sample Sets by Combining Unsupervised Meta-Training with CNNs

    Yuxiong Wang, Martial Hebert

    NeurIPS, 2016. [PDF][Poster]


    Unsupervised Binary Codes

    Few-Shot Hash Learning for Image Retrieval

    Yuxiong Wang, Liangke Gui, Martial Hebert

    ICCV Workshops, 2017. [PDF]

    Discovering Unsupervised Binary Codes for Learning from Small Sample Sets

    Yuxiong Wang, Martial Hebert

    CVPR BigVision Workshop, 2016. [Poster]


    Transfer Learning from Unsupervised Universal Sources

    Model Recommendation: Generating Object Detectors from Few Samples

    Yuxiong Wang, Martial Hebert

    CVPR, 2015. [PDF][Poster][Project]

Non-Negative Matrix and Tensor Factorization and Its Applications

    Non-Negative Matrix Factorization: A Comprehensive Review

    Non-Negative Matrix Factorization: A Comprehensive Review

    Yuxiong Wang, Yu-Jin Zhang

    IEEE Transactions on Knowledge and Data Engineering, vol. 25, no. 6, pp.1336-1353, 2013. [PDF]


    Non-negative Matrix Factorization: A Comprehensive Review

    Image Inpainting via Weighted Sparse Non-Negative Matrix Factorization

    Yuxiong Wang, Yu-Jin Zhang

    ICIP, 2011. [PDF][Poster]


    Neighborhood Preserving Non-Negative Tensor Factorization for Image Representation

    Yuxiong Wang, Liangyan Gui, Yu-Jin Zhang

    Oral Presentation, ICASSP, 2012. [PDF]


    Non-Negative Matrix Factorization for Image Representation

    Yuxiong Wang

    Best Master’s Thesis, Tsinghua University, May 2012.

Sparse Representation and Dictionary Learning for Image Classification

    Self-Explanatory Sparse Representation for Image Classification

    Self-Explanatory Sparse Representation for Image Classification

    Yuxiong Wang*, Baodi Liu*, Bin Shen, Yu-Jin Zhang, Martial Hebert

    ECCV, 2014 (* indicates equal contribution). [PDF][Poster]


    Learning Dictionary on Manifolds for Image Classification

    Learning Dictionary on Manifolds for Image Classification

    Bao-Di Liu, Yuxiong Wang, Yu-Jin Zhang, Bin Shen

    Pattern Recognition, vol. 46, no. 7, pp. 1879-1890, 2013. [PDF]


    Blockwise Coordinate Descent Schemes for Sparse Representation

    Bao-Di Liu, Yuxiong Wang, Bin Shen, Yu-Jin Zhang, Yanjiang Wang

    ICASSP, 2014. [PDF]


    Discriminant Sparse Coding for Image Classification

    Baodi Liu, Yuxiong Wang, Yu-Jin Zhang, Yin Zheng

    ICASSP, 2012. [PDF]

Object Tracking

    Mean-Shift Object Tracking through 4-D Scale Space

    Yuxiong Wang, Yu-Jin Zhang, Xiaohua Wang

    Journal of Electronics and Information Technology, vol. 32, no. 7, pp.1626-1632, 2010. [PDF]


    Video Based Object Tracking

    Yuxiong Wang

    Best Bachelor Thesis, Beijing Institute of Technology, June 2009.