黄文炳

Structures Matter

I am now a senior researcher at Tencent AI Lab. I got my Ph.D. degree of computer science from Tsinghua University in July 2017, before which I studied Mathematics in Beihang University. My current research mainly lies in the areas of machine learning and computer vision, with particular focus on developing models/methods that analyze structured data (usually Non-Euclidean and Non-Grid like) by, for example, characterizing invariance and equivalence among them. Those types of data that I have coped with include but are not limited to:

  • Graphs: graph representation learning, graph neural networks;
  • Videos: motion representation learning, event captioning, image2video translation, action classification and detection, imitation learning from videos;
  • Manifolds: infinite Grassmannian (modeled by linear dynamical systems);

Contact me by [hwenbing at 126 dot com].


News

[2019-07-23] 2 papers accepted by ICCV-2019 (one oral and one poster) !

[2019-07-10] I was invited as a session chair of IJCAI-2019, see you in Macao!

[2019-07-10] 2 papers accepted by ACMMM-2019!

[2019-06-20] 2 papers accepted by TIP!

[2019-04-22] 1 paper accepted by ICML-2019!

[2019-02-25] 1 paper accepted by CVPR-2019!




Selected Publications
        
        Wenbing Huang, Tong Zhang, Yu Rong, Junzhou Huang. 
        Adaptive Sampling Towards Fast Graph Representation Learning
        Annual Conference on Neural Information Processing Systems (NeurIPS), Montreal, Canada, 2018.
         paper  code  poster

        Xuguang Duan*, Wenbing Huang*, Chuang Gan, Jingdong Wang, Wenwu Zhu, Junzhou Huang. 
        Weakly Supervised Dense Event Captioning in Videos
        Annual Conference on Neural Information Processing Systems (NeurIPS), Montreal, Canada, 2018. 
         paper code

        Lijie Fan*, Wenbing Huang*, Chuang Gan, Stefano Ermon, Boqing Gong, Junzhou Huang
        End-to-End Learning of Motion Representation for Video Understanding,  
        IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, USA, 2018. (Spotlight)  
        paper code

        Wenbing Huang, Mehrtash Harandi, Tong Zhang, Lijie Fan, Fuchun Sun, Junzhou Huang. 
        Efficient Optimization for Linear Dynamical Systems with Applications to Clustering and Sparse Coding
        Annual Conference on Neural Information Processing Systems (NeurIPS), Long Beach, USA, 2017.
        paper  code

        Wenbing Huang, Fuchun Sun, Lele Cao, Deli zhao, Huaping Liu, Mehrtash Harandi. 
        Sparse Coding and Dictionary Learning with Linear Dynamical Systems
        IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, Nevada, USA, 2016. (Oral
        paper code

        (* Equal)


Full Publications
  • 2019
        Yu Rong, Wenbing Huang, Tingyang Xu, Junzhou Huang,
        The Truly Deep Graph Convolutional Networksfor Node Classification,
        arXiv, 2019

        Guangyao Shen, Wenbing Huang, Chuang Gan, Mingkui Tan, Junzhou Huang, Wenwu Zhu, Boqing Gong,
        Facial Image-to-Video Translation by A Hidden Affine Transformation,
        ACM International Conference on Multimedia (ACMMM), 2019.

        Xuguang Duan, Qi Wu, Chuang Gan, Yiwei Zhang, Wenbing Huang, Anton van den Hengel and Wenwu Zhu,
        Watch, Reason and Code: Learning to Represent Videos Using Program,
        ACM International Conference on Multimedia (ACMMM), 2019.

        Runhao Zeng, Chuang Gan, Peihao Chen, Wenbing Huang, Qingyao Wu, Mingkui Tan,
        Breaking Winner-takes-all: Iterative-winners-out Networks for Weakly Supervised Temporal Action Localization,
        IEEE Transactions on Image Processing (TIP), 2019.
        paper

        Fuchun Sun*, Tao Kong*, Wenbing Huang, Chuanqi Tan, Bin Fang, Huaping Liu,
        Feature Pyramid Reconfiguration with Consistent Loss for Object Detection,
        IEEE Transactions on Image Processing (TIP), 2019. 
         paper

        Tong Zhang, Pan Ji, Mehrtash Harandi, Wenbing Huang, Hongdong Li, 
        Neural Collaborative Subspace Clustering,
        International Conference on Machine Learning (ICML), 2019.
          paper
        
        Chaoqi Chen*, Weiping Xie*, Wenbing Huang, Yu Rong, Xinghao Ding, Yue Huang, Tingyang Xu, Junzhou Huang,
        Progressive Feature Alignment for Unsupervised Domain Adaptation, 
        IEEE Conference on Computer Vision and Pattern Recognition (CVPR),  2019.
          paper

        Jia Li, Yu Rong, Hong Cheng, Helen Meng, Wenbing Huang, Junzhou Huang,
        Semi-Supervised Graph Classification: A Hierarchical Graph Perspective,   
        International World Wide Web Conference (WWW),  San Francisco, USA, 2019.
          paper

        Lijie Fan, Wenbing Huang, Chuang Gan, Junzhou Huang, Boqing Gong
        Controllable Image-to-Video Translation: A Case Study on Facial Expression Generation
        The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI),  Honolulu, Hawaii, USA, 2019. (Oral)
        paper

        Xiangpeng LiJingkuan SongLianli GaoXianglong Liu, Wenbing Huang, Xiangnan He, Chuang Gan
        Beyond RNNs: Positional Self-Attention with Co-Attention for Video Question Answering
        The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI),  Honolulu, Hawaii, USA, 2019. (Oral)

        Mingxuan Jing*, Xiaojian Ma*Wenbing Huang Fuchun Sun, Huaping Liu
        Task Transfer by Preference-Based Cost Learning
        The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI),  Honolulu, Hawaii, USA, 2019. 
         paper

  • 2018
        Xuguang Duan*, Wenbing Huang*, Chuang Gan, Jingdong Wang, Wenwu Zhu, Junzhou Huang. 
        Weakly Supervised Dense Event Captioning in Videos
        Annual Conference on Neural Information Processing Systems (NeurIPS), Montreal, Canada, 2018. 
         paper code

        Wenbing Huang, Tong Zhang, Yu Rong, Junzhou Huang. 
        Adaptive Sampling Towards Fast Graph Representation Learning
        Annual Conference on Neural Information Processing Systems (NeurIPS), Montreal, Canada, 2018.
         paper  code  poster

        Wenbing Huang*, Lijie Fan*, Mehrtash Harandi, Lin Ma, Huaping Liu, Wei Liu, Chuang Gan
        Towards Efficient Action Recognition: Principal Backpropagation for Training Two-Stream Networks.
        IEEE Transactions on Image Processing (TIP), 2018. 
         paper

        Lijie Fan*, Wenbing Huang*, Chuang Gan, Stefano Ermon, Boqing Gong, Junzhou Huang
        End-to-End Learning of Motion Representation for Video Understanding,  
        IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, USA, 2018. (Spotlight oral)  
        paper code

        Tao Kong, Fuchun SunWenbing Huang, Huaping Liu. 
        Deep Feature Pyramid Reconfiguration for Object Detection
        European Conference on Computer Vision (ECCV), Munich, Germany, 2018.  
         paper

        Chunfang Liu, Wenbing Huang, Fuchun Sun, Minnan Luo, Chuanqi Tan. 
        LDS-FCM: A Linear Dynamical System-based Fuzzy C-Means Method for Tactile Recognition
        IEEE Transactions on Fuzzy Systems (TFS), 2018.
         paper
  • 2017
        Wenbing Huang, Mehrtash Harandi, Tong Zhang, Lijie Fan, Fuchun Sun, Junzhou Huang. 
        Efficient Optimization for Linear Dynamical Systems with Applications to Clustering and Sparse Coding
        Annual Conference on Neural Information Processing Systems (NeurIPS), Long Beach, USA, 2017.
        paper  appendix code
  • 2016
        Wenbing Huang, Fuchun Sun, Lele Cao, Mehrtash Harandi. 
        Analysing Linear Dynamical Systems: from Modeling to Coding and Learning, 2016. 
        paper code

        Wenbing Huang, Fuchun Sun, Lele Cao, Deli zhao, Huaping Liu, Mehrtash Harandi. 
        Sparse Coding and Dictionary Learning with Linear Dynamical Systems
        IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, Nevada, USA, 2016. (Oral
        paper code

        Wenbing Huang, Lele Cao, Fuchun Sun, Deli zhao, Huaping Liu, Shanshan Yu. 
        Learning Stable Linear Dynamical Systems with the Weighted Least Square Method
        International Joint Conference on Artificial Intelligence (IJCAI), New York, USA, 2016. 
        paper code

        LeleCao, Kotagiri Ramamohanarao, FuchunSun, Hongbo Li, Wenbing Huang, Zay Maung Aye. 
        Efficient Spatio-Temporal Tactile Object Recognition with Randomized Tiling Convolutional Networks in a Hierarchical Fusion Strategy
        AAAI Conference on Artificial Intelligence (AAAI), Phoenix, Arizona, USA, 2016.
        paper

  • 2015
        Fuchun Sun, Chunfang Liu, Wenbing Huang, Jianwei Zhang, 
        Object Recognition and Grasp Planning using Visual and Tactile Sensing
        IEEE Transactions on Systems, Man and Cybernetics: Systems, 2015.

        Wenbing Huang, Deli Zhao, Fuchun Sun, Huaping Liu, Edward Chang. 
        Scalable Gaussian Process Regression Using Deep Neural Networks.     
        International Joint Conference on Artificial Intelligence (IJCAI), Buenos Aires, Argentina, 2015. 
         paper code

Codes and Tools

        AS-GCN: A package for fast graph representation learning. The distributed version will come soon. Link to code .
        
        TVNet: A package for end-to-end optical flow estimation. Link to code.
        
        WSDEC: The code for Weakly Supervised Dense Event Captioning in Videos. Link to code
    
        LDS: A tool box for linear dynamical systems modeling. Link to code



Co-supervised students

      Lijie Fan (phd, MIT)
        Xuguang Duan (phd, THU)
        Runhao Zeng (phd, SCUT)
        Guangyao Shen (phd, THU)
        Chao Yang (phdTHU)
        Mingxuan Jing (phdTHU)
        Xiaojian Ma (Undergraduate studentTHU)