sAI Zhang

Research Scientist, Meta Reality Labs

 Incoming Assistant Professor, New York University 

About Me

I am working as a research scientist at Meta Reality Labs. Previously I was a postdoctoral fellow at Harvard University, and I received my PhD degree of Computer Science at the same university. Before coming to Harvard, I received my bachelor's degree and master's degree in Electrical Engineering and Statistics from University of Toronto.

 

My chinese name is 张赛骞.

 

I am incredibly excited to announce that I will be joining New York University in Fall 2024 as a tenure-track Assistant Professor of Electrical Engineering and Computer Science! I am looking for highly motivated PhD students and visiting students/researchers. Interested candidates are strongly encouraged to contact me by email, together with resume and transcripts.

Research Overview

Overall, my research interest lies in algorithm/hardware codesign for efficient deep neural network (DNN) implementation.

Application & Algorithm:

  • Efficient DNN computing, pruning, quantization, NAS

  • Recent interest: parameter efficient finetuning for LLM, efficient self-supervised learning, privacy for AI

Hardware architecture:

  • Domain-specific accelerator for compute-intensive AI applications, New compute paradigm for DNN

  • Recent interest: AI accelerator for on-device transfer learning and contrastive learning

Others:

  • Multi-agent reinforcement learning and its application

  • Federated learning

  • AI compiler

News

[11/2023] Serving as TPC for DAC'24 and ISQED'24.

[10/2023] Our paper "Co-Designing AI Models and DRAMs for On-Device Training" is accepted by HPCA 2024!

[9/2023] Our paper on efficient reinforcement learning, which I co-authored with my high school mentee, Gavin An, has been accepted for publication in JEI.

[6/2023] I gave two talks on DNN hardware and algorithm codesign at Tsinghua University and Peking University.

[5/2023] Our paper "Co-Designing AI Models and DRAMs for On-Device Training" is submitted to Arxiv. This paper proposes an algorithm/hardware codesign solution for efficient on-chip transfer learning which completely eliminates the off-chip DRAM traffic during the training process.

[3/2023] My high school mentee, Gavin An, has successfully finished his AI project on efficient reinforcement learning. A paper is submitted to JEI.

[9/2022] Start working at Meta!

[7/2022] Our paper “Hyperspherical Federated Learning" is accepted by ECCV 2022!

[7/2022] I gave an invited talk at AI times on multi-agent reinformcent learning and its applications.

[6/2022] I gave an invited talk at IEEE Dallas Circuits and Systems Conference (DCAS), 2022.

[4/2022] I started my AI memtorship at Veritas AI!

[2/2022] I started my postdoc study at Harvard!

[12/2021] I successfully defended my PhD!

[11/2021] Our paper “Learning Advanced Client Selection Strategy for Federated Learning" is accepted by AAAI 2022!

[10/2021] Our paper “FAST: DNN Training Under Variable Precision Block Floating Point with Stochastic Rounding" is accepted by IEEE HPCA 2022!

[06/2021] Finished my internship at Microsoft, such a great place for research!

[03/2021] I started my virtual internship at Microsoft Research, Redmond.

[03/2021] I gave a guest lecture on DNN accelerator design on Havard Course ES201, hosted by Prof. Demba Ba.

[01/2021] One paper got accepted by IEEE International Symposium on Circuits & Systems (ISCAS), 2021.

[12/2020] I presented (virtually) our work "Succinct and Robust Multi-Agent Communication With Temporal Message Control" in NeurIPS 2020.

[11/2020] I presented (virtually) our work "Term quantization: furthering quantization at run time" in SC 2020.

[11/2020] Our paper “Training for Multi-resolution Inference Using Reusable Quantization Terms" is accepted by ACM ASPLOS 2021!

[09/2020] Our paper "Succinct and Robust Multi-Agent Communication With Temporal Message Control" is accepted by NeurIPS 2020!

[08/2020] I presented our work "Adaptive Distributed Convolutional Neural Network Inference at the Network Edge with ADCNN" in ICPP 2020.

[06/2020] Our paper "Term Revealing: Furthering Quantization at Run Time on Quantized DNNs" is accepted by ACM/IEEE SC 2020!

[05/2020] One paper "Adaptive Distributed Convolutional Neural Network Inference at the Network Edge with ADCNN" is accepted by ACM ICPP 2020!

[02/2020] One paper is accepted by IEEE Symposium on Security and Privacy (S&P) Deep Learning and Security workshop, 2020.

[12/2019] I attended NeurIPS and presented our work in Vancouver, Canada.

[11/2019] Our paper "RTN: Reparameterized Ternary Network" is accepted by AAAI 2020!