Mi Zhang's Profile Picture


Mi Zhang

Associate Professor
Director, AIoT and Machine Learning Systems Lab
Department of Computer Science and Engineering
The Ohio State University

Email: mizhang.1 [at] osu [dot] edu
Office: 495 Dreese Labs, 2015 Neil Ave, Columbus, OH 43210



Welcome

I am an Associate Professor and the Director of AIoT and Machine Learning Systems Lab in the Department of Computer Science and Engineering at The Ohio State University.

Mission Statement
: The proliferation of edge devices such as mobile phones and Internet of of Things (IoT) as well as the gigantic amount of data they generate make it no longer feasible to transmit all the data to the cloud for processing. Such constraint fuels the need to move the intelligence from the cloud to the edge where data reside. The key mission of my lab is Empowering Billions of Everyday Devices with AI to realize the vision of Artificial Intelligence of Things (AIoT). To achieve this mission, my group focuses on its core challenges related to sensing, intelligence, connectivity, efficiency, automation, scalability, privacy, inclusiveness and important applications such as healthcare and sustainability. Achieving such goal requires a combination of approaches. My work draws insights from a broad set of disciplines including mobile & embedded systems, AI/machine learning, wireless networking, distributed systems, and human-centered computing.

Research Topics
: Specifically, my students and I work on the following topics:

  • Edge AI for mobile, AR, wearables, IoT as well as networked and distributed edge-cloud systems.
  • Large Language Models, Generative AI, and their domain-specific applications.
  • AI for wireless networking, video/audio analytics, and smart sensing systems.
  • Systems for AI
  • Mobile Health and Human-Centered AI for applications such as digital mental health, smart hearing aids, sign language translation, pill identification.
  • Federated Learning

My group also closely collaborates with industry (e.g., Google Research, Meta (Facebook), Amazon, and Microsoft Research) to make practical impact.

At OSU, I lead and co-organize the OSU AIoT seminar series with the goal to push the frontier of AIoT forward.

Short Bio
: I received my B.S. in Electrical Engineering from Peking University (PKU) in Beijing, China. I received my Ph.D. in Computer Engineering and M.S. in both Electrical Engineering and Computer Science from University of Southern California (USC). I spent one year as a Postdoctoral Associate in Computing and Information Science at Cornell University. Before joining OSU, I was an Assistant Professor and then a tenured Associate Professor at Michigan State University.

I have received a number of awards for my work. I am the 4th Place Winner (1st Place in U.S. and Canada) of 2019 Google MicroNet Challenge (CIFAR-100 Track), the Third Place Winner of 2017 NSF Hearables Challenge, and the champion of 2016 NIH Pill Image Recognition Challenge. I am the recipient of seven best paper awards and nominations. I am also the recipient of NSF CRII Award, Facebook Faculty Research Award, Amazon Machine Learning Research Award, MSU Innovation of the Year Award, and the inaugural USC ECE SIPI Distinguished Alumni Award in the Junior/Academia category for my contributions to mobile/edge computing in my early career.

Some Recent Talks
:

  • Building Efficient, Scalable, and Inclusive Federated Learning Systems. Meta (Facebook), 2022.
  • Empowering the Next Billion Devices with Deep Learning. Stanford Machine Learning Systems (MLSys) Seminar, 2021. [Video]
  • Keynote: Encoding is an Important Design Decision in Neural Architecture Search. ICML'21 AutoML Workshop.
  • Efficient On-Device Distributed Deep Learning via Importance Sampling. Facebook AI Systems Faculty Summit, 2020.




Selected Awards


News

  • 03/2023:
    Invited to attend the research summit for egocentric perception hosted by Meta Reality Lab. The future of AR glasses is bright!
  • 03/2023:
    Congratulations Zhongwei for landing his internship at Tencent AI Lab!
  • 03/2023:
    Congratulations Xin for landing his internship at Microsoft Research Asia!
  • 03/2023:
    Congratulations Zhongwei for being invited to give a talk at Stanford MedAI Seminar Series!
  • 02/2023:
    Congratulations Xin for being awarded the highly competitive OSU College Allocated Fellowship!
  • 01/2024:
    Recent advancements in Generative AI have enabled a new wave of AI revolution. The implications of such advancements for Internet of Things (IoT) are profound. In this article, we share our views on the applications, challenges, and opportunities of IoT in the era of Generative AI.
  • 12/2023:
    After months of hard work, Efficient Large Language Models: A Survey is online.
  • 09/2023:
    Thanks NSF for the NeTS Medium grant to support our project on high-performing LoRa with edge AI.
  • 02/2023:
    Feel extremely honored and excited to receive the inaugural USC ECE SIPI Distinguished Alumni Award in the Junior/Academia category for my contributions to mobile/edge computing in my early career.
  • 01/2023:
    Congratulations Samiul for being awarded the highly competitive OSU College Allocated Fellowship!
  • 12/2022:
    Thanks Meta Reality Labs for the generous faculty award for supporting our research!
  • 09/2022:
    FedRolex, our work on model-heterogeneous federated learning is accepted to NeurIPS'22.
  • 08/2022:
    After eight amazing years at MSU, our group has joined Department of Computer Science and Engineering at OSU.
  • 08/2022:
    Congratulations Shen for denfending his PhD thesis! Wish you all the best at Google Research!
  • 07/2022:
    Invited to give a talk at Meta (Facebook) to talk about how we build efficient, scalable, and inclusive federated learning systems.
  • 04/2022:
    Invited to give a talk at Hong Kong University of Science and Technology (HKUST).
  • 04/2022:
    Invited to give a talk at the Efficient AI Seminar at Rutgers University.
  • 03/2022:
    Our vision paper on enabling federated learning on potentially billions of IoT devices is published at IEEE Internet of Things Magazine (IEEE IoTM).
  • 01/2022:
    Congratulations Shen for landing his summer internship at Google Brain!
  • 01/2022:
    Deep AutoAugment, our AutoML work on deep data augmentation policy search is accepted to ICLR'22.
  • 01/2022:
    Invited to serve as the area chair of the International Conference on Automated Machine Learning (AutoML'22).
  • 01/2022:
    PyramidFL (Client Selection for Federated Learning) is accepted to ACM MobiCom'22.
  • 12/2021:
    We are organizing the 1st International Workshop on Federated Learning for Computer Vision (FedVision) at CVPR'22. Please consider submitting your best work.
  • 12/2021:
    Invited to give a talk at the USC Center for Cyber-Physical Systems and Internet of Things Seminar.
  • 11/2021:
    NELoRa won the Best Paper Award at ACM SenSys'21, and is selected as ACM SIGMOBILE Research Highlight.
  • 11/2021:
    Invited to attend this year's Google Workshop on Federated Learning and Analytics.
  • 10/2021:
    Invited to give a talk at the Stanford Machine Learning Systems (MLSys) Seminar.
  • 09/2021:
    Mercury (On-Device Distributed DNN Training), FedMask (Federated Learning), and NELoRa (AI-enhanced LoRa) are accepted to ACM SenSys'21.
  • 08/2021:
    Invited to give a talk at the 7th Workshop on Energy Efficient Machine Learning and Cognitive Computing.
  • 07/2021:
    Invited to serve as the program committee of the Conference on Machine Learning and Systems (MLSys'22).
  • 07/2021:
    In collaboration with many colleagues, we are very excited to publish A Field Guide to Federated Optimization, a paper that serves as a guide on federated learning research.
  • 05/2021:
    Congratulations Xiao for denfending his PhD thesis! Wish you all the best at Amazon!
  • 05/2021:
    CATE, our AutoML work on neural architecture search based on computation-aware neural architecture encoding is accepted as long talk (top 3%) at ICML'21.
  • 05/2021:
    Congratulations Shen and Yu for landing their internships at Google Research and Amazon!
  • 04/2021:
    Invited to give a Keynote at ICML'21 AutoML Workshop.
  • 04/2021:
    Our book 创新工场讲AI课:从知识到实践 (Chinese) is published.
  • 12/2020:
    FedML won the Best Paper Award at NeurIPS'20 Federated Learning Workshop.
  • 10/2020:
    Invited to attend and give a talk at the Facebook AI Systems Faculty Summit.
  • 09/2020:
    arch2vec, our AutoML work on neural architecture search based on unsupervised architecture representation learning is accepted to NeurIPS'20.
  • 09/2020:
    Distream, our distributed deep learning framework for large-scale video analytics is accepted to ACM SenSys'20.
  • 09/2020:
    WiSIA (Wi-Fi See It All) is accepted to ACM SenSys'20.
  • 08/2020:
    Invited to give a tutorial on AutoML at ECCV'20 Tutorial session: From HPO to NAS: Automated Deep Learning.
  • 08/2020:
    FlexDNN, our on-device deep learning framework for efficient mobile vision is accepted to ACM/IEEE SEC'20 and is the Best Paper Award Nominee.
  • 07/2020:
    In collaboration with many colleagues, we are very excited to introduce FedML, a research library and benchmark for federated learning.
  • 07/2020:
    Invited to attend the Workshop on Federated Learning and Analytics hosted by Google.
  • 07/2020:
    MutualNet is accepted as oral (top 2%) at ECCV'20. MutualNet is an adaptive ConvNet that is able to achieve adaptive accuracy-efficiency trade-offs at runtime for on-device AI.
  • 05/2020:
    Congratulations Xiao, Yu, and Shen for landing their internships at Amazon and ByteDance!
  • 04/2020:
    Thanks Amazon AWS AI for the AWS Machine Learning Research Award!
  • 03/2020:
    SecWIR is accepted to ACM MobiSys'20.
  • 02/2020:
    Thanks Facebook Research for the Facebook Faculty Research Award on Systems for Machine Learning!
  • 02/2020:
    Our AI-enabled smart hearing aid receives the 2020 MSU Innovation of the Year Award!
  • 12/2019:
    Congratulations Biyi for denfending his PhD thesis! Wish you all the best at Microsoft!
  • 12/2019:
    We are the 4th Place Winner (1st Place in U.S. and Canada) of the Google MicroNet Challenge CIFAR-100 Track hosted at NeurIPS'19! Official Announcement. This is our third global competition win over the past 4 years. We have made our algorithm open source, and hope it can push the research area of on-device/edge AI forward.
  • 08/2019:
    HM-NAS, our AutoML work on weight-sharing based neural architecture search won the Best Paper Award Nominee at ICCV'19 Neural Architects Workshop.
  • 05/2019:
    Congratulations Shen for landing his internship at Bosch Research!
  • 08/2018:
    Thanks NSF for the NeTS Small grant (Co-PI) to fund our Mobile Internet of Things (Mobile IoT) project!
  • 07/2018:
    NestDNN, our on-device deep learning framework that enables resource-aware multi-tenant on-device AI is accepted to ACM MobiCom'18.
  • 05/2018:
    The Dark Side of Operational Wi-Fi Calling Services won the Best Paper Award at IEEE CNS'18. The work reported in this paper also won the Google Security Reward.
  • 05/2018:
    Congratulations Biyi for landing his internship at Microsoft!
  • 04/2018:
    Our vision paper on realizing ubiquitous mixed reality by combining Internet of Things (IoT) and mixed reality (MR) is published at ACM SIGMOBILE GetMobile. Our ACM UbiComp'18 paper is one concrete realization of this vision.
  • 12/2017:
    Invited to serve on the Technical Advisory Board of the UCLA Depression Grand Challenge.
  • 09/2017:
    Interviewed by WIRED to share my view on NVIDIA's Deep Learning Accelerator and on accelerating deep learning on mobile and embedded devices.
  • 08/2017:
    We are the Third Place Winner of the NSF Hearables Challenge! Official NSF Announcement. We are invited by NSF to present our work at ACM UbiComp'17. Media coverage: ACM TechNews, TUN, R&D Magazine, AAU, MSU Today.
  • 07/2017:
    DeepASL, our deep learning based American sign language (ASL) translation system that enables ubiquitous and non-intrusive ASL translation at both word and sentence levels is accepted to ACM SenSys'17. Media coverage: NSF (video), MSU (video), NVIDIA, NPR (radio interview), Smithsonian, AAU, Futurity, MSU Today.
  • 05/2017:
    Our review paper on personal sensing and machine learning for digital mental health is published at the Annual Review of Clinical Psychology (Impact Factor: 12.214).
  • 04/2017:
    Honored to be selected as the 2017 NIH Mobile Health (mHealth) Scholar!
  • 02/2017:
    MobileDeepPill, our award-winning on-device deep learning based mobile pill recognition system is accepted to ACM MobiSys'17.
  • 09/2016:
    Thanks NSF for the PFI:BIC grant (PI)! We are very grateful for receiving this grant to develop personal sensing technologies and mobile sensor data analytics techniques to combat depression on university campuses. Media coverage: NSF (video), Smithsonian Magazine, MSU Today, EdTech, iTechPost, etc.
  • 08/2016:
    We are the First Place Winner of the NIH Pill Image Recognition Challenge! Official NIH Announcement | Media coverage: MSU Today, ABC News (TV), etc.
  • 08/2016:
    Thanks NSF for the CSR Small grant (PI)!
  • 08/2016:
    Honored to receive the NSF CRII Award! Media coverage: MSU Today, ACM TechNews, NPR, ABC News (TV), etc.
  • 05/2016:
    Congratulations Biyi for landing his internship at Intel Labs!
  • 05/2016:
    AirSense, our AIoT system for indoor air quality sensing and analytics is accepted to ACM UbiComp'16. Media coverage: The Atlantic, MSU Today, Futurity, etc.
  • 02/2016:
    BodyScan, our wireless sensing system for contactless activity and vital sign monitoring is accepted to ACM MobiSys'16.
  • 01/2016:
    HeadScan, our wireless sensing system for contactless activity monitoring is accepted to ACM/IEEE IPSN'16. Media coverage: Stanford Medicine, Fox News (TV interview), ReadWrite, Futurity, MSU Today, Medgadget.
  • 09/2015:
    Congratulations Biyi for landing his internship at Bell Labs!
  • 07/2015:
    DoppleSleep, our device-free wireless sensing system for contactless sleep monitoring won the Best Paper Honorable Mention Award at ACM UbiComp'15. Media coverage: MIT Technology Review, etc.
  • 07/2015:
    MyBehavior, our reinforcement learning-based mobile recommendation system is accepted to ACM UbiComp'15. Media coverage: MIT Technology Review, Mashable, MobiHealth News, etc.
  • 07/2015:
    Our paper on personal sensing for depression detection using mobile phone sensors and machine learning is accepted to JMIR. It is one of the JMIR All-Time Top Article now. See the rank here. Media coverage: TIME, CNN, TechCrunch, The Verge, CBS News, Fox News, Discovery News, Daily Mail, The Times, Newsweek, Mirror, The Telegraph, The Washington Post, The Huffington Post, Los Angeles Times, Chicago Tribune, Futurity, WebMD, US News, etc.
  • 06/2014:
    BodyBeat, our mobile sensing system that listens to sounds inside human body for continuous health monitoring is accepted to ACM MobiSys'14, and is selected as ACM SIGMOBILE Research Highlight. Media coverage: MIT Tech Review, Wall Street Journal, New Scientist.


Sponsors

Our research is generously sponsored by the following federal agencies and industry partners. We express our sincere gratitude to their support.

Sponsors




© Mi Zhang. All rights reserved.