Le Yang (杨乐)

Assistant Professor
School of Information and Communications Engineering, Xi'an Jiaotong University
Email: yangle15@xjtu.edu.cn

|Google Scholar| |Github|

Address: Room 140, Pengkang building, Xingqing Campus.
               Room 8058, #4 building, Chuangxingang Campus.

Research Interests

  • Deep learning
  • Dynamic neural networks / Adaptive inference
  • Video understanding
  • Efficient deep architectures
  • Collaborative intelligence

    •       NOTE: We are looking for self-motivated Master candidates. Please drop me an email with your resume if you are interested in our group. You can also contact Prof. Fan Li and Prof. Lijun He for Master and PhD candidates in our group.

    Research Experience

  • 12/2021 - 07/2022, Visiting Scholar, ML Group, Aalto University.    Advised by Prof. Arno Solin.

    • Education Background

  • 09/2015 - 06/2021, PhD, Department of Automation, Tsinghua University.    Advised by Prof. Shiji Song and Prof. Gao Huang.
  • 09/2011 - 07/2015, B.E., Department of Automation, Northwestern Polytechnical University.    (GPA Top 1/120)

    • Selected Publications

    * equal contribution. + corresponding author.

    Dynamic Spatial Focus for Efficient Compressed Video Action Recognition. [paper]
    Ziwei Zheng, Le Yang+, Yulin Wang, Miao Zhang, Lijun He, Gao Huang, Fan Li.
    IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT) 2023.
  • We propose the dynamic spatial focus for compressed video recognition. It is the first dynamic neural network for videos in compressed format (such as MPEG4 and HEVC). The adaptive patch selection strategy crops out the irrelevant motion noise in motion vectors, as well as reduce the spatial redundancy of the inputs, leading to the high efficiency of our method in the compressed domain.


    • CondenseNet V2: Sparse Feature Reactivation for Deep Networks. [paper][code][知乎]
    Le Yang*, Haojun Jiang*, Ruojin Cai, Yulin Wang, Shiji Song, Gao Huang+, Qi Tian.
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2021.
  • We propose a new feature reusing method in deep networks through dense connectivity, which can simultaneously learn to 1) selectively reuse a set of most important features from preceding layers; and 2) actively update a set of preceding features to increase their utility for later layers.


    • Dynamic neural networks: A survey. [paper]
    Yizeng Han, Gao Huang+, Shiji Song, Le Yang, Honghui Wang, Yulin Wang.
    IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI) 2021.
  • In this survey, we comprehensively review the rapidly developing area, dynamic neural networks. The important research problems, e.g., architecture design, decision making scheme, and optimization technique, are reviewed systematically. We also discuss the open problems in this field together with interesting future research directions.


    • Resolution Adaptive Networks for Efficient Inference. [paper][code]
    Le Yang*, Yizeng Han*, Xi Chen*, Shiji Song, Jifeng Dai, Gao Huang+.
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2020.
  • The proposed Resolution Adaptive Network (RANet) makes use of spatial redundancy in images to conduct the adaptive inference for the first time. The RANet is inspired by the intuition that low-resolution representations are sufficient for classifying “easy” inputs containing large objects with prototypical features, while only some “hard” samples need spatially detailed information.
    • Revisiting Locally Supervised Learning: an Alternative to End-to-end Training. [paper][code]
    Yulin Wang, Zanlin Ni, Shiji Song, Le Yang, Gao Huang+.
    International Conference on Learning Representations (ICLR) 2021.
  • By revisiting the locally supervised learning, we experimentally show that simply training local modules with E2E loss tends to collapse task-relevant information at early layers, and hence hurts the performance of the full model. To avoid this issue, we propose an information propagation (InfoPro) loss, which encourages local modules to preserve as much useful information as possible, while progressively discard task-irrelevant information. As InfoPro loss is difficult to compute in its original form, we derive a feasible upper bound as a surrogate optimization objective, yielding a simple but effective algorithm.


    • Glance and Focus: a Dynamic Approach to Reducing Spatial Redundancy in Image Classification. [paper] [code]
    Yulin Wang, Kangchen Lv, Rui Huang, Shiji Song, Le Yang, Gao Huang+.
    Neural Information Processing Systems (NeurIPS) 2020.
  • Inspired by the fact that not all regions in an image are task-relevant, we propose a novel framework that performs efficient image classification by processing a sequence of relatively small inputs, which are strategically selected from the original image with reinforcement learning. Such a dynamic decision process naturally facilitates adaptive inference at test time, i.e., it can be terminated once the model is sufficiently confident about its prediction and thus avoids further redundant computation.

    • Publication List

    * equal contribution. + corresponding author.

    [C1] Le Yang, Shiji Song+ and C. L. Philip Chen. "Transductive Transfer Learning Based on Broad Learning System," IEEE International Conference on Systems, Man, and Cybernetics, 2018.

    [J1] Le Yang, Shiji Song+, Yanshang Gong, Gao Huang and Cheng Wu, "Nonparametric Dimension Reduction via Maximizing Pairwise Separation Probability," IEEE Transactions on Neural Networks and Learning Systems (T-NNLS), 2019.
    [J2] Yiming Chen, Shiji Song+, Shuang Li, Le Yang and Cheng Wu, "Domain Space Transfer Extreme Learning Machine for Domain Adaptation," IEEE Transactions on Cybernetics (T-Cyber), 2019.

    [C2] Le Yang*, Yizeng Han*, Xi Chen*, Shiji Song, Jifeng Dai and Gao Huang+, "Resolution Adaptive Networks for Efficient Inference," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020. (CCF-A)
    [C3] Yulin Wang, Kangchen Lv, Rui Huang, Shiji Song, Le Yang and Gao Huang+, "Glance and Focus: a Dynamic Approach to Reducing Spatial Redundancy in Image Classification," Advances in Neural Information Processing Systems (NeurIPS), 2020. (CCF-A)

    [C4] Le Yang*, Haojun Jiang*, Ruojin Cai, Yulin Wang, Shiji Song, Gao Huang+ and Qi Tian, "CondenseNet V2: Sparse Feature Reactivation for Deep Networks," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021. (CCF-A)
    [C5] Yulin Wang, Zanlin Ni, Shiji Song, Le Yang and Gao Huang+, "Revisiting Locally Supervised Learning: an Alternative to End­-to-­end Training," International Conference on Learning Representations (ICLR), 2021.
    [C6] Le Yang, Xiaoli Gong, Yizeng Han, Lijun He and Fan Li+, "Dark-channel mixed attention based neural networks for smoke detection in fog environment," Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2021 ACM International Symposium on Wearable Computers (UbiComp/ISWC), 2021. (CCF-A)
    [C7] Ziwei Zheng, Le Yang, Liejun Wang and Fan Li+, "AD-DARTS: Adaptive Dropout for Differentiable Architecture Search," the First CAAI International Conference (CICAI), 2021.
    [J3] Le Yang, Shiji Song+, Shuang Li, Yiming Chen and C. L. Philip Chen, "Discriminative Dimension Reduction via Maximin Separation Probability Analysis," in IEEE Transactions on Cybernetics (T-Cyber), 2021.
    [J4] Le Yang, Shiji Song, Shuang Li+, Yiming Chen and Gao Huang, "Graph Embedding-Based Dimension Reduction With Extreme Learning Machine," in IEEE Transactions on Systems, Man, and Cybernetics: Systems (T-SMC-A), 2021.
    [J5] Yizeng Han, Gao Huang+, Shiji Song, Le Yang, Honghui Wang and Yulin Wang. "Dynamic neural networks: A survey," IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2021. (CCF-A)
    [J6] Gao Huang+, Chunjiang Ge, Tianyu Xiong, Shiji Song, Le Yang, Baoxian Liu, Wenjun Yin and Cheng Wu, "Large scale air pollution prediction with deep convolutional networks," Journal of SCIENCE CHINA Information Sciences, 2021. (CCF-A)
    [J7] Le Yang, Yiming Chen, Shiji Song, Fan Li and Gao Huang, "Deep siamese networks based change detection with remote sensing images," Remote Sensing, 2021.
    [J8]Yizeng Han, Gao Huang+, Shiji Song, Le Yang, Yitian Zhang and Haojun Jiang, "Spatially adaptive feature refinement for efficient inference," IEEE Transactions on Image Processing (T-IP), 2021. (CCF-A)

    [C8] Le Yang, Zelin Yang, Ziwei Zheng, Lijun He, Fan Li+ and C.L. Philip Chen, "Anomaly Detection based on Broad Leaning System for Rolling Element Bearing Fault Diagnosis", Adjunct Proceedings of the 2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2022 ACM International Symposium on Wearable Computers (UbiComp/ISWC), 2022. (CCF-A)
    [C9]Zixi Wang, Yuan Zhang, Le Yang, Fan Li+, "Privacy-preserved Intermediate Feature Compression for Cyber-Physical Systems," Adjunct Proceedings of the 2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2022 ACM International Symposium on Wearable Computers (UbiComp/ISWC), 2022. (CCF-A)

    [J9] Le Yang, Ziwei Zheng, Jian Wang, Shiji Song, Gao Huang and Fan Li+, "AdaDet: An Adaptive Object Detection System based on Early-exit Neural Networks," IEEE Transactions on Cognitive and Developmental Systems (T-CDS), 2023.
    [J10] Le Yang, Zelin Yang, Shiji Song, Fan Li+ and C.L. Philip Chen, "Twin Broad Learning System for Fault Diagnosis of Rotating Machinery," IEEE Transactions on Instrumentation and Measurement (T-IM), 2023.
    [J11] Ziwei Zheng, Le Yang+, Yulin Wang, Miao Zhang, Lijun He, Gao Huang and Fan Li, "Dynamic Spatial Focus for Efficient Compressed Video Action Recognition," IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), 2023.

    Selected Honors

    • The 2021 Postdoctoral Innovative Talent Program, advised by CAS Fellow, Prof. Xiaohong Guan.
    • Beijing Outstanding Graduate Award, 2021 (Highest honor for graduate set by the government of Beijing).
    • China National Scholarship for Graduate, Ministry of Education of China, 2019. (Highest level of scholarship for graduate set by the government of China).
    • Outstanding Undergraduate Student, 2015 at NWPU.
    • China National Scholarship for Undergraduate, Ministry of Education of China, 2014 & 2013 & 2012. (Highest level of scholarship for undergraduate set by the government of China)

    Professional Activities

    • Technical Programm Committee of UbiComp 2023.
    • Program Committee (PC) member of IJCAI 2021.
    • Reviewer for IJCV, T-PAMI, T-NNLS, T-Cyber, T-CSVT, ...
    • Reviewer for CVPR, ICCV, IJCAI, NeurIPS, ICML, ICLR...