Personalized Federated Learning with Moreau Envelopes
Canh T. Dinh, Nguyen H. Tran, Tuan Dung Nguyen, (2020). "Personalized Federated Learning with Moreau Envelopes." Neurips 2020).
CV
Education
- B.S. in Electronics and Telecommunication, Hanoi University of Science and Technology, Ha Noi, Viet Nam, 2010-2015
- M.S. in Data Science at University of Grenoble, Grenoble, France, 2019
- Ph.D in Computer Science at The University of Sydney Sydney (USYD), Australia, 2019-2022
Research experience
- The School of Computer Science, USYD (Fed 2019 – Sep 2022): Phd Candidate
- The University of Sydney
- Duties included: Design novel distributed algorithms to improve privacy, efficiency, and effectiveness of Federated Learning System and apply “Multi-model” approaches including Meta-Learning, Multi-task learning, and Personalization to exploit cross-client statistical correlation in Federated Learning.
- Supervisor: Dr. Nguyen H. Tran (https://nguyenhoangtran.github.io/)
- Integrated circuit Design Lab: Research Assistant
- Hanoi University of Science and Technology
- Duties included: Research to find out and deploy new compression algorithms for Motion Estimation block including IME (integer motion estimation) and Intra Prediction in HEVC (High Efficiency Video Coding) on FPGA.
- Supervisor: Dr. Nguyen Vu Thang
Work experience
- Canva (Dec 2021 – Current): Machine Learning Engineer
- Develop image to text models using Transformer to automatically generate titles for templates designed by Canva.
- Develop image/text to text models to automatically generate keywords for templates designed by Canva.
- Ingenico Vietnam Company Limited (Feb 2018 – Aug 2018): Software Engineer
- Develop applications for POS devices include chip cards, contactless, and magstripe transactions for MasterCard, Amex, JCB, Visa
- Humax R&D Center (Dec 2015 – Oct 2017): Software Engineer
- Develop software applications for automotive products
- VTTEK - Viettel Network Technologies Center (Jun 2015 – Dec 2015): Software Engineer
- Building Volib libraries for Packet loss enhancement, Tone Detection (DTMF) and Audio codec libs for (G711, AMR, AMR-WB, GSMFR, GSMHR) to integrate into the Media-Gateway project which uses DSP C6000 multicore processors to process (encode, decode audio) audio in UMTS system
Skills
- Technical Skills: Machine Learning, Distributed systems, High-Performance Computing, Embedded Systems
- Programming Languages: Python, C++, Java, Verilog, VHDL
- Platforms and Tools: PyTorch, TensorFlow, Caffe2, Apache Hadoop, Git, SVN
Publications
Federated Learning with Proximal Stochastic Variance Reduced Gradient Algorithms
Canh T. Dinh, Nguyen H. Tran, Tuan Dung Nguyen, Wei Bao, Albert Y. Zomaya, Bing B. Zhou, (2020). "Federated Learning with Proximal Stochastic Variance Reduced Gradient Algorithms." International Conference on Parallel Processing (ICPP)-2020.
Federated Learning over Wireless Networks: Convergence Analysis and Resource Allocation
Canh T. Dinh, Nguyen H. Tran, Minh N. H. Nguyen, Choong Seon Hong, Wei Bao, Albert Y. Zomaya, and Vincent Gramoli, (2020). "Federated Learning over Wireless Networks: Convergence Analysis and Resource Allocation." IEEE/ACM Transactions on Networking.
DONE: Distributed Approximate Newton-type Method for Federated Edge Learning
Canh T. Dinh, Nguyen H. Tran, Tuan Dung Nguyen, Amir Rezaei Balef, Wei Bao, Bing B. Zhou, and Albert Y. Zomaya, (2022). "DONE: Distributed Approximate Newton-type Method for Federated Edge Learning." IEEE Transactions on Parallel and Distributed Systems.
A New Look and Convergence Rate of Federated Multi-Task Learning with Laplacian Regularization
Canh T. Dinh, Tung T. Vu, Nguyen H. Tran, Minh N. Dao, Hongyu Zhang, (2022). "A New Look and Convergence Rate of Federated Multi-Task Learning with Laplacian Regularization." Under review (minor revision) IEEE Transactions on Neural Networks and Learning Systems.