Fei (Travis) Dai
About Me
Hello and welcome! I’m Fei (Travis) Dai, a passionate educator and researcher serving as a lecturer at the Eastern Institute of Technology. With a deep-rooted fascination for the interplay between light and computation, my academic and professional journey revolves around harnessing optical technologies to revolutionize how we process information and communicate.
At the core of my research interests lie optical interconnect systems—a vital component for the next generation of high-speed, energy-efficient computing. I am deeply immersed in exploring parallel distributed computation of deep neural networks (DNN), aiming to unlock new computational speed and efficiency horizons. My work on optical DNN accelerators is driven by the vision of creating groundbreaking hardware to keep pace with the ever-growing demands of artificial intelligence and machine learning.
Communication optimization within these systems is another area where I focus my efforts, striving to enhance how our devices and infrastructures converse in the digital age. Integrating embedded systems and the Internet of Things (IoT) into this framework forms the backbone of my approach towards creating a seamlessly connected world where information flows as freely and efficiently as light.
I am constantly inspired by the potential of optical technologies to redefine our computational limits and am committed to advancing this field through rigorous research, innovative teaching, and collaborative exploration. Whether you’re a student eager to delve into the wonders of optical computing, a fellow researcher keen on discussing potential collaborations, or simply curious about the future of technology, I invite you to join me on this exciting journey.
Research Interests (Cross-discipline includes:)
- Optical Interconnects
- DNN Training Acceleration
- IoT and Embedded Systems
News
- [Jan. 2024] Our paper about the all-gather operation has been accepted by the IEEE Open Journal of the Communications Society (OJ-COMS).
- [Nov. 2023] Our paper about distributed DNN training performance comparison is accepted to ICA3PP 2023.
- [Oct. 2023] Our paper about the all-reduce algorithm has been accepted to The 52nd International Conference on Parallel Processing (ICPP) 2023.
- [Sept. 2023] Our paper about the all-gather operation in optical interconnect systems is accepted to ICC 2023.
- [Mar. 2023] Our paper about FCNN training performance comparison is accepted by The Journal of Supercomputing.
- [Feb. 2023] Our paper about distributed DNN training has been accepted to PPoPP 2023.
Selected Publications
-
ICPP
The 52nd International Conference on Parallel Processing (ICPP), 2023. -
OJ-COMS
IEEE Open Journal of the Communications Society (OJ-COMS), 2024.
Services
Guest Editor
Conference Reviewers
International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT) 2020-2021 2019 International Symposium on Parallel Architectures, Algorithms, and Programming (PAAP) 2019 International Workshop on Programming Models and Applications for Multicores and Manycores (PMAM 2023)
Journal Reviewers
IEEE Transactions on Parallel and Distributed Systems (TPDS) PLOS ONE ICT Express Journal of Supercomputing (SUPE) Microprocessors and Microsystems IEEE ACCESS Journal of Optics and Photonics Research
Teaching
Electronics and Internet of Things
Automation and Embedded Systems
Powered by Jekyll and Minimal Light theme.