About me
I am currently the co-founder of FunAmi AI, a company dedicated to promoting inclusive education through artificial intelligence (AI). Previously, I worked as an algorithm engineer specializing in E-commerce search at Meituan (美团). My research covers a wide range of topics, including computer vision and deep learning-based model compression. Throughout my career, I have been committed to advancing autonomous driving through visual perception. Then, I shifted my focus to generative search utilizing large language models (LLM) for various reasons.
I worked on high-performance computing for deep neural networks on edge devices during my internship at Maicro. This included the development of neural network inference frameworks on ARM chips and the algorithm verification of the Quantized Neural Process Unit (QNPU).
I earned my Ph.D. from the Institute of Automation, Chinese Academy of Sciences (CASIA) in 2022, under the esteemed guidance of Prof. Jian Cheng. I received my B.S. degree from the Beijing University of Posts and Telecommunications (BUPT) in 2017.
Awards
- Outstanding Reviewer, CVPR 2024
- Recognized among the top 2% of reviewers
- Top Reviewer, NeurIPS 2023
- Ranked among the top 10.2% of reviewers (1,197 out of 11,725)
- Top Reviewer, NeurIPS 2022
- Ranked among the top 10.7% of reviewers (1,101 out of 10,334); invited as session chair
- Outstanding Reviewer, ICML 2022
- Recognized among the top 10% of reviewers; invited as session chair
- Winner, MicroNet: Large-Scale Model Compression Competition, ImageNet Track
- Organized by Google Research, DeepMind, OpenAI, Facebook @ NeurIPS 2019
- Winner, MicroNet: Large-Scale Model Compression Competition, CIFAR Track
- Organized by Google Research, DeepMind, OpenAI, Facebook @ NeurIPS 2019
- Winner, Challenge on Learned ISP
- Organized by ETH, Google, Adobe, Nvidia, Huawei, Qualcomm @ ECCV AIM (Advances in Image Manipulation) 2020
- Runner-up, Efficient Super-Resolution Challenge
- Organized by ETH, Google, Adobe, Nvidia, Huawei, Qualcomm @ ECCV AIM (Advances in Image Manipulation) 2020
- Runner-up, Lightweight Face Recognition Challenge & Workshop, DeepGlint Light Track
- Organized by DeepGlint, iQIYI, Imperial College London @ ICCV 2019
Authored Book
- I am honored to have participated in the writing of the book “深度神经网络高效计算” by the CLab team. After more than two years of preparation, the book has finally been published.
Selected Papers
- Z* : Zero-shot Style Transfer via Attention Reweighting
- Yingying Deng*, Xiangyu He*, Fan Tang, Weiming Dong
- Computer Vision and Pattern Recognition (CVPR), 2024
- Code, Co-First Author
- Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning
- Yanpeng Sun*, Qiang Chen*, Xiangyu He*, Jian Wang, Haocheng Feng, Junyu Han, Errui Ding, Jian Cheng, Zechao Li, Jingdong Wang
- Conference on Neural Information Processing Systems (NeurIPS), 2022
- Code, Co-First Author
- Toward Accurate Binarized Neural Networks With Sparsity for Mobile Application
- Peisong Wang*, Xiangyu He*, Jian Cheng
- IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2022
- Code, Co-First Author
- Revisiting L1 Loss in Super-Resolution: A Probabilistic View and Beyond
- Xiangyu He, Jian Cheng
- Tech report, January 2022
- Towards Fully Sparse Training: Information Restoration with Spatial Similarity
- Weixiang Xu, Xiangyu He, Ke Cheng, Peisong Wang, Jian Cheng
- Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2022
- Code
- APRIL: Finding the Achilles’ Heel on Privacy for Vision Transformers
- Jiahao Lu, Xi Sheryl Zhang, Tianli Zhao, Xiangyu He, Jian Cheng
- Computer Vision and Pattern Recognition (CVPR), 2022
- Optimization-based Post-training Quantization with Bit-split and Stitching
- Peisong Wang, Weihan Chen, Xiangyu He, Qiang Chen, Qingshan Liu, Jian Cheng
- IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2022
- Code
- Generative Zero-shot Network Quantization
- Xiangyu He, Jiahao Lu, Weixiang Xu, Qinghao Hu, Peisong Wang, Jian Cheng
- Conference on Computer Vision and Pattern Recognition Workshop (CVPRw), 2021
- Dynamic Dual Gating Neural Networks
- Fanrong Li, Gang Li, Xiangyu He, Jian Cheng
- International Conference on Computer Vision (ICCV), 2021
- Code, Oral
- Skeleton-based Action Recognition with Shift Graph Convolutional Network
- Ke Cheng, Yifan Zhang, Xiangyu He, Weihan Chen, Jian Cheng, Hanqing Lu
- Computer Vision and Pattern Recognition (CVPR), 2020
- Code, Oral
- Towards Accurate Post-training Network Quantization via Bit-split and Stitching
- Peisong Wang, Qiang Chen, Xiangyu He, Jian Cheng
- International Conference on Machine Learning (ICML), 2020
- Code
- Unsupervised Network Quantization via Fixed-Point Factorization
- Peisong Wang, Xiangyu He, Qiang Chen, Anda Cheng, Qingshan Liu, Jian Cheng
- IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2020
- Code
- Sparsity-inducing Binarized Neural Networks
- Peisong Wang*, Xiangyu He*, Gang Li, Tianli Zhao, Jian Cheng
- Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2020
- Code, Spotlight
- Soft Threshold Ternary Networks
- Weixiang Xu, Xiangyu He, Tianli Zhao, Qinghao Hu, Peisong Wang, Jian Cheng
- International Joint Conference on Artificial Intelligence (IJCAI), 2020
- Code
- Proxybnn: Learning Binarized Neural Networks via Proxy Matrices
- Xiangyu He, Zitao Mo, Ke Cheng, Weixiang Xu, Qinghao Hu, Peisong Wang, Qingshan Liu, Jian Cheng
- European Conference on Computer Vision (ECCV), 2020
- Spotlight
- Location-aware Upsampling for Semantic Segmentation
- Xiangyu He, Zitao Mo, Qiang Chen, Anda Cheng, Peisong Wang, Jian Cheng
- Tech report, November 2019
- Code
- Compact Global Descriptor for Neural Networks
- Xiangyu He, Ke Cheng, Qiang Chen, Qinghao Hu, Peisong Wang, Jian Cheng
- Tech report, June 2019
- Code
- Ode-inspired Network Design for Single Image Super-resolution
- Xiangyu He*, Zitao Mo*, Peisong Wang, Yang Liu, Mingyuan Yang, Jian Cheng
- Computer Vision and Pattern Recognition (CVPR), 2019
- Code
- K-nearest Neighbors Hashing
- Xiangyu He, Peisong Wang, Jian Cheng
- Computer Vision and Pattern Recognition (CVPR), 2019
- Code
- Bitstream: Efficient Computing Architecture for Real-time Low-power Inference of Binary Neural Networks on CPUs
- Tianli Zhao, Xiangyu He, Jian Cheng, Jing Hu
- ACM International Conference on Multimedia (ACM MM), 2018
- Learning Compression from Limited Unlabeled Data
- Xiangyu He, Jian Cheng
- European Conference on Computer Vision (ECCV), 2018
- Code
Honor
- Outstanding New Employee Award of the Autonomous Vehicle Delivery Department, Meituan 2023
- National Scholarship, CASIA 2020
Academic Activity
- Reviewer: CVPR, ECCV, ICCV, ICML, NeurIPS, ICLR, AAAI, WACV, ACCV, IJCV, TIP