Publications
Peer-Reviewed Conference Publications
(2019 – present; selected)
[Google Scholar] [Patents]
[Industrial R&D Experience] [Technical Reports] [Leetcode]
Potential Publication Venues:
[Machine Learning] ICML, NeurIPS, ICLR;
[ML/AI Applications] CVPR/ECCV/ICCV, ACL/EMNLP, AAAI/IJCAI;
[Distributed/Cloud/Mobile/Database Systems] NSDI, OSDI, SOSP, SC, VLDB, MobSys, MobiCom, Sensys, HotOS, EuroSys, ICDCS, etc.
- [27] PhD Thesis – Federated and Distributed Machine Learning at Scale: From Systems to Algorithms to Applications
Chaoyang He, University of Southern California
[overleaf]
TDLR: My PhD thesis is an initial version of my company FedML Inc., a startup building open and collaborative AI anywhere at any scale. See our latest progress at https://fedml.ai. During My Ph.D., I did a full stack of scientific publications in ML/FL algorithms, distributed systems, security/Privacy, AI applications, and visionary impacts for FedML.ai, the federated learning and Edge AI Platform. All my papers are to make it better. Check the following video to see how easy and funny we do machine learning socially, privately, and collaboratively. After my PhD graduation, I will continue my research in the real world through FedML Inc.
- [26] LightSecAgg: a Lightweight and Versatile Design for Secure Aggregation in Federated Learning
Jinhyun So*, Chaoyang He*, Chien-Sheng Yang*, Songze Li, Yu Qian, Salman Avestimehr
(* means co-1st authors; I contributed to ML+Sys co-design)
Accepted to MLSys 2022 – Fifth Conference on Machine Learning and Systems
TDLR: this is a fun team work in our lab. It makes FedML.ai more secure! - [25] Partial Model Averaging in Federated Learning: Performance Guarantees and Benefits
Sunwoo Lee, Anit Kumar Sahu, Chaoyang He, Salman Avestimehr
Accepted to International Workshop on Trustable, Verifiable and Auditable Federated Learning in Conjunction with AAAI 2022 (Oral Presentation)
[Arxiv] - [24] SSFL: Tackling Label Deficiency in Federated Learning via Personalized Self-Supervision
Chaoyang He, Zhengyu Yang, Erum Mushtaq, Sunwoo Lee, Mahdi Soltanolkotabi, Salman Avestimehr
Accepted to International Workshop on Trustable, Verifiable and Auditable Federated Learning in Conjunction with AAAI 2022 (Best Paper Award)
[Arxiv] - [23] SPIDER: Searching Personalized Neural Architecture for Federated Learning
Erum Mushtaq, Chaoyang He, Jie Ding, Salman Avestimehr
Accepted to International Workshop on Trustable, Verifiable and Auditable Federated Learning in Conjunction with AAAI 2022
[Arxiv] - [22] FedCV: A Federated Learning Framework for Diverse Computer Vision Tasks
Chaoyang He, Alay Dilipbhai Shah, Zhenheng Tang, Di Fan, Adarshan Naiynar Sivashunmugam, Keerti Bhogaraju, Mita Shimpi, Li Shen, Xiaowen Chu, Mahdi Soltanolkotabi, Salman Avestimehr
Accepted to International Workshop on Trustable, Verifiable and Auditable Federated Learning in Conjunction with AAAI 2022 (Oral Presentation)
[Arxiv] - [21] SpreadGNN: Serverless Multi-task Federated Learning for Molecular Graphs
Chaoyang He*, Emir Ceyani*, Keshav Balasubramanian*, Murali Annavaram, Salman Avestimehr
Accepted to AAAI 2022 (Thirty-Sixth AAAI Conference on Artificial Intelligence)
A short version has been accepted to International Workshop on Federated Learning for User Privacy and Data Confidentiality in Conjunction with ICML 2021 (FL-ICML’21) and Deep Learning on Graphs: Method and Applications with KDD 2021 (DLG-KDD’21)
[Arxiv] - [20] AutoCTS: Automated Correlated Time Series Forecasting
Xinle Wu, Dalin Zhang, Chenjuan Guo, Chaoyang He, Bin Yang, Christian Jensen
Accepted to 48th International Conference on Very Large Data Bases, Sydney, Australia – September 05-09, 2022 (VLDB 2022)
[Arxiv]
TDLR: neural architecture search (NAS) for time-series forecasting. - [19] Federated Learning for Internet of Things: Applications, Challenges, and Opportunities
Tuo Zhang, Lei Gao, Chaoyang He, Mi Zhang, Bhaskar Krishnamachari, Salman Avestimehr
[Arxiv], Under Review, 2021 - [18] FedNLP: Benchmarking Federated Learning Methods for Natural Language Processing Tasks
Bill Yuchen Lin*, Chaoyang He*, Zihang Zeng, Hulin Wang, Yufen Huang, Mahdi Soltanolkotabi, Xiang Ren, Salman Avestimehr
(* means co-1st authors)
[Arxiv]
Accepted to NAACL 2022 - [17] LightSecAgg: a Lightweight and Versatile Design for Secure Aggregation in Federated Learning
Chien-Sheng Yang, Jinhyun So, Chaoyang He, Songze Li, Yu Qian, Salman Avestimehr
Accepted to 2021 IEEE Information Theory Workshop (ITW 2021)
[Arxiv] - [16] OmniLytics: A Blockchain-based Secure Data Market for Decentralized Machine Learning
Jiacheng Liang, Songze Li, Wensi Jiang, Bochuan Cao, Chaoyang He
Accepted to International Workshop on Federated Learning for User Privacy and Data Confidentiality in Conjunction with ICML 2021 (FL-ICML’21)
[Arxiv] [ICML official website] - [15] FairFed: Enabling Group Fairness in Federated Learning
Yahya Ezzeldin, Shen Yan, Chaoyang He, Emilio Ferrara, Salman Avestimehr
Accepted to AAAI 2022
A short version has been accepted 2021 Neural Information Processing Systems Workshop “New Frontiers in Federated Learning: Privacy, Fairness, Robustness, Personalization and Data Ownership” (NeurIPS 2021)
[Arxiv] - [14] Layer-wise Adaptive Model Aggregation for Scalable Federated Learning
Sunwoo Lee, Tuo Zhang, Chaoyang He, Salman Avestimehr
Arxiv, Under Review, 2021
- [13] Federated Learning for Internet of Things: A Federated Learning Framework for On-device Anomaly Data Detection
Tuo Zhang*, Chaoyang He*, Tianhao Ma, Lei Gao, Mark Ma, Salman Avestimehr
(* means co-1st authors)
accepted to ACM Embedded Networked Sensor Systems SenSys 2021 (AIChallengeIoT)
[Proceeding] [Arxiv] [GitHub]
TLDR: IoT x Federated Learning, from FedML.ai - [12] A Field Guide to Federated Optimization (with Google and CMU). 2021
Jianyu Wang, Zachary Charles, Zheng Xu, Gauri Joshi, H. Brendan McMahan, Blaise Aguera y Arcas, Maruan Al-Shedivat, Galen Andrew, Salman Avestimehr, Katharine Daly, Deepesh Data, Suhas Diggavi, Hubert Eichner, Advait Gadhikar, Zachary Garrett, Antonious M. Girgis, Filip Hanzely, Andrew Hard, Chaoyang He, Samuel Horvath, Zhouyuan Huo, Alex Ingerman, Martin Jaggi, Tara Javidi, Peter Kairouz, Satyen Kale, Sai Praneeth Karimireddy, Jakub Konecny, Sanmi Koyejo, Tian Li, Luyang Liu, Mehryar Mohri, Hang Qi, Sashank J. Reddi, Peter Richtarik, Karan Singhal, Virginia Smith, Mahdi Soltanolkotabi, Weikang Song, Ananda Theertha Suresh, Sebastian U. Stich, Ameet Talwalkar, Hongyi Wang, Blake Woodworth, Shanshan Wu, Felix X. Yu, Honglin Yuan, Manzil Zaheer, Mi Zhang, Tong Zhang, Chunxiang Zheng, Chen Zhu, Wennan Zhu
TDLR: want to learn distributed/federated optimization? This is a good starting point!
- [11] MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge
Geng Yuan, Xiaolong Ma, Wei Niu, Zhengang Li, Zhenglun Kong, Ning Liu, Yifan Gong, Zheng Zhan, Chaoyang He, Qing Jin, Siyue Wang, Minghai Qin, Bin Ren, Yanzhi Wang, Sijia Liu, Xue Lin
Accepted to Conference on Neural Information Processing Systems, NeurIPS 2021
TDLR: edge training is unrealistic? We try to make it happen! - [10] PipeTransformer: Automated Elastic Pipelining for Distributed Training of Transformers
Chaoyang He, Shen Li (Facebook AI Research), Mahdi Soltanolkotabi, Salman Avestimehr
Accepted to ICML 2021 (International Conference on Machine Learning 2021)
[Arxiv] [Proceeding] [Homepage] [Slides] [Animation] [Open Source Code] [DistML.ai]
TDLR: PipeTransformer is featured by PyTorch.org official website: https://pytorch.org/blog/pipetransformer-automated-elastic-pipelining/ - [9] FedGraphNN: A Federated Learning System and Benchmark for Graph Neural Networks
Chaoyang He*, Keshav Balasubramanian*, Emir Ceyani*, Carl Yang, Han Xie, Lichao Sun, Lifang He, Liangwei Yang, Philip S. Yu, Yu Rong, Peilin Zhao, Junzhou Huang, Murali Annavaram, Salman Avestimehr
(* equal contribution)
A short version has been accepted to DPML@ICLR 2021 and GNNSys@MLSys 2021
[ICLR 2021 Workshop Video Presentation] [MLSys 2021 Workshop Camera Ready Version] [MLSys 2021 Poster] [Code]
TDLR: in the past 5 years, graph learning is the most popular subarea in data mining. We extend it to federated/distributed training! Our co-authors have published two NeurIPS 2021 papers with the help of our library!
- [8] FedML: A Research Library and Benchmark for Federated Machine Learning
Chaoyang He, Songze Li (HKUST, Stanford), Jinhyun So (USC), Mi Zhang (USC), Hongyi Wang (U Wisconsin Madison), Xiaoyang Wang (UIUC), Praneeth Vepakomma (MIT), Abhishek Singh (MIT), Hang Qiu (Stanford), Li Shen (Tencent), Peilin Zhao (Tencent), Yan Kang (WeBank), Yang Liu (WeBank), Ramesh Raskar (MIT), Qiang Yang (HKUST, WeBank), Murali Annavaram, Salman Avestimehr
The short version of FedML white paper won Best Paper Award at SpicyFL@NeurIPS 2020
[BibTex] [Homepage] [Arxiv] [Slack] [Documentation] [Video] [Slides] [Best Paper Award] [Code]
TDLR: FedML Ecosystem (https://FedML.ai) is a family of open research libraries to facilitate federated learning research in diverse application domains. It includes FedML Core Framework, FedNLP (Natural Language Processing), FedCV (Computer Vision), FedGraphNN (Graph Neural Networks), FedIoT (Internet of Things), and FedMobile (Smartphones). With the fundamental support from FedML Core Framework, FedML Ecosystem enables three computing paradigms: on-device training for edge devices (cross-device FL), distributed computing (cross-silo FL), and single-machine simulation, and also promotes diverse algorithmic research with flexible and generic API design and comprehensive reference baseline implementations (federated optimizer, private/security algorithms, models, and datasets). Compared with TFF and LEAF, FedNLP and FedCV greatly enrich the diversity of data sets and learning tasks. FedNLP supports various popular task formulations in the NLP domain, such as text classification, sequence tagging, question answering, seq2seq generation, and language modeling. FedCV can help researchers evaluate the three most representative tasks: image classification, image segmentation, and object detection. Moreover, FedGraphNN is the first FL research platform for analyzing graph-structured data using Graph Neural Networks in a distributed computing manner, filling the gap between federated learning and the data mining field. FedGraphNN collects, preprocess, and partitions 36 datasets from 7 domains such as molecule ML (drug discovery, bioinformatics, etc.), Social networks, Recommendation Systems, and Knowledge Graphs. Going beyond traditional AI applications, FedIoT and FedMobile further extend FL to perform in wireless communication (e.g., 5G) and mobile computing (e.g., embedded IoT devices such as Raspberry PI, smartphones running on Android OS). All in all, FedML Ecosystem aims to provide a one-stop scientific research platform through FedML Ecosystem and finally realize trustworthy ML/AI, which is more secure, scalable, efficient, and ubiquitous.
- [7] Group Knowledge Transfer: Federated Learning of Large CNNs at the Edge
Chaoyang He, Murali Annavaram, Salman Avestimehr
Accepted to NeurIPS 2020 (Conference on Neural Information Processing Systems 2020)
[BibTex] [Arxiv] [NeurIPS Official Site Video Presentation] [Gather Town] [Project Page] [Poster] [Slides] [Code]
TDLR: going beyond transfer data in centralized learning and transfer model/gradient in federated learning, we proposed a new direction that transfer “knowledge” to do collaborative AI. - [6] Towards non-I.I.D. and invisible data with FedNAS: Federated Deep Learning via Neural Architecture Search
Chaoyang He, Murali Annavaram, Salman Avestimehr
Accepted to CVPR 2020 Workshop on Neural Architecture Search and Beyond for Representation Learning, 2020
[BibTex] [Arxiv] [video] [Code]
TDLR: AutoML x Federated Learning - [5] MiLeNAS: Efficient Neural Architecture Search via Mixed-Level Reformulation
Chaoyang He*, Haishan Ye*, Li Shen, Tong Zhang (HKUST, Tencent AI Lab)
Accepted to CVPR 2020 (IEEE/CVF Conference on Computer Vision and Pattern Recognition 2020)
[BibTex] [arxiv] [video] [Code]
TDLR: Thanks for the guidance of Professor Tong Zhang, Director of Tencent AI Lab.
- [4] Advances and Open Problems in Federated Learning (with Googlers). 2019
Peter Kairouz, H Brendan McMahan, Brendan Avent, Aurélien Bellet, Mehdi Bennis, Arjun Nitin Bhagoji, Keith Bonawitz, Zachary Charles, Graham Cormode, Rachel Cummings, Rafael GL d’Oliveira, Salim El Rouayheb, David Evans, Josh Gardner, Zachary Garrett, Adria Gascón, Badih Ghazi, Phillip B Gibbons, Marco Gruteser, Zaid Harchaoui, Chaoyang He, Lie He, Zhouyuan Huo, Ben Hutchinson, Justin Hsu, Martin Jaggi, Tara Javidi, Gauri Joshi, Mikhail Khodak, Jakub Konečný, Aleksandra Korolova, Farinaz Koushanfar, Sanmi Koyejo, Tancrede Lepoint, Yang Liu, Prateek Mittal, Mehryar Mohri, Richard Nock, Ayfer Özgür, Rasmus Pagh, Mariana Raykova, Hang Qi, Daniel Ramage, Ramesh Raskar, Dawn Song, Weikang Song, Sebastian U Stich, Ziteng Sun, Ananda Theertha Suresh, Florian Tramèr, Praneeth Vepakomma, Jianyu Wang, Li Xiong, Zheng Xu, Qiang Yang, Felix X Yu, Han Yu, Sen Zhao
[BibTex] [Arxiv] [Jeff Dean’s Tweet] [Publication at Foundations and Trends in Machine Learning]
Published at FnTML 2020 (Foundations and Trends in Machine Learning, the chief editor is Michael Jordan)
TDLR: 105 pages, 485 references, 22 Googlers, and 36 academics at 24 institutions!
- [3] Central Server Free Federated Learning over Single-sided Trust Social Networks
Chaoyang He, Conghui Tan (WeBank), Hanlin Tang, Shuang Qiu, Ji Liu (Kwai AI Lab, Professor at University of Rochester)
Preprint, 2020
A short version has been accepted to SpicyFL@NeurIPS 2020 (Lightning Talk)
[BibTex] [Arxiv] [Lightning Talk] [Poster] [Code]
TDLR: Decentralized Federated Learning! - [2] Cascade-BGNN: Toward Efficient Self-supervised Representation Learning on Large-scale Bipartite Graphs
Chaoyang He, Tian Xie, Yu Rong, Wenbing Huang, Junzhou Huang, Xiang Ren, Cyrus Shahabi
[Arxiv] - [1] Collecting Indicators of Compromise from Unstructured Text of Cybersecurity Articles using Neural-Based Sequence Labelling
Zi Long, Lianzhi Tan, Shengping Zhou, Chaoyang He, Xin Liu
International Joint Conference on Neural Networks, IJCNN 2019
[Arxiv]
TDLR: natural language processing, security; work done when I was at Tencent, the company that developed WeChat…
Industrial R&D Experience
I was a full-stack software engineer and team manager. I have lots of experience in successful Internet product publication. My experience spans nearly all Internet-related software systems development, including cloud computing, distributed systems, machine learning systems, applied machine learning, operating system, and mobile applications. Business domain: cloud computing, automotive intelligent system, autonomous driving, smartphone, Maps/LBS, games, etc
- Tencent Cloud. 2017-2018
Worked on business-oriented cloud computing solution; led a team to develop the 1st generation cloud computing solution for the automotive industry.

- Tencent Venus Distributed Machine Learning System. 2016-2017
Worked with Shengping Zhou (now CTO@AlphaCloud), developing the large-scale distributed machine learning system for MIG, Tencent. This platform supports various machine learning algorithms and models, including, LR, SVM, GBDT, DNN, etc. One more thing: I have to highlight that one module of Tencent Venus contains the first practice of federated learning system in China, even in the world, earlier than Google FL. At Tencent, we supported vertical federated learning infrastructure for WeBank to model Tencent social network features into their financial models.

- Tencent Location Data Computing Platform. 2016-2017
worked on streaming data mining systems and algorithms; this platform contains billion-level daily users; I learned the entire system architecture and related real-time data analytics algorithms.

- Speech Recognition and Natural Language Processing for Tencent Automotive Operating System. 2015-2016
This project inspired me to explore ML models, algorithms, and systems; deep learning is very fancy that year; we developed a deep learning model compression engine for our product (at that time, TensorFlow Lite is not released yet)

- Tencent Games: a Pokemon Go-like Mobile Game Engine 2016
worked on Unity3D and C++ -based game engine; this is a cross-group collaboration, very fun; I learned a lot of engineering and production culture at Tencent Game.

- Tencent AI in Car (also called Tencent Automotive Service). 2014-2016
I was leading an operating system team; worked on the embedded operating system (Android/QNX) and related back-end cloud services, including more than 10 applications and system core service, such as WeChat MyCar, Tencent Maps, Tencent Music, Tencent Video, etc. Please watch the video below to have a taste how Chinese Internet companies advertise their products:-)

- As a Team Leader, I Developed the 1st generation of Smartphone Navigation SDK, Application, and Service in China (Baidu Navigation SDK). 2012-2013
Baidu Navigation SDK, Application, and Service is a cross-platform (C++/Java/Objective-C) engine that requires computer graphics (OpenGL ES), routing algorithms (A*, CH), offline and online data compiler and networking service (Hadoop; C++/PHP backend), complex navigation UI message state control and dispatch, cross-platform, layered, and modularized complex system design. I am familiar with all of these core modules.

- As a Software Engineer, I Contributed to the First Generation of Huawei Smartphone. 2011-2012
The picture below shows Huawei Ascend P6 (Emotion UI)

Technical Report
My industrial experience spans cloud computing, distributed system, applied machine learning, data computing platform, mobile computing (Android/iOS/IoT devices), and their applications in AI, Speech Recognition, Games, Maps, and Navigation.
- 2018 – Data-Driven Cloud Computing Platform and Machine Learning System for the Internet of Vehicles
- 2016 – The Large-Scale Distributed System and Real-Time Location Data Mining
- 2017 – TAI: An Intelligent Operating System and Open Platform
- 2016 – Speech Recognition System for Connected and Autonomous Driving Car
- 2016 – WeGameMap: Real World Map Rendering Engine for Location-Based Mobile Game
- 2015 – WeLink: a High-Performance Vehicle-Mobile Networking Library
- 2014 – TMAP: A System Framework for Mobile Maps and Navigation
- 2016 – Efficient Spatial Anti-Aliasing Rendering for Line Joins on Vector Maps
- 2014 – WeCross: a Mobile and Vehicle C/C++ Cross-Platform Library
- 2017 – A High Precision Private Car Trajectory Dataset and an Open Source Location Data Computing Platform
- 2017 – A Open Source High Reliable Location SDK for the Automotive Industry
- 2013 – The First Generation Mobile Navigation SDK Design in China
- 2012 – Map State Switching under Multi-trigger Condition Based on Finite State Machine Design Pattern
- 2012 – A General Downloader Engine for iOS/Android Platform
- 2012 – The Best Practice for Java Native Interface (JNI) Development on Android Operating Platform
- 2014 – Taxi-Calling Platform and the Open Source Code
- 2015 – A UX design of Bluetooth Steering Wheel Wireless Controller and the communication system design
- 2016 – Sliding Window Algorithm Template to Solve All the Leetcode Substring Search Problem.
Leetcode
An Algorithmic Source Code Template for Solving Many Substring Search Problems, 2016
(1.8K upvote, 2000+ stars, 167.8K views)