Amirhossein Reisizadeh

Amirhossein Reisizadeh Ph.D. candidate Department of Electrical and Computer Engineering UC Santa Barbara Contact: reisizadeh[at]ucsb[dot]edu Google Scholar page

Short Bio

I am a Ph.D. candidate in ECE department at UC Santa Barbara working with Prof. Ramtin Pedarsani. I have been fortunate to collaborate with Prof. Salman Avestimehr, Prof. Hamed Hassani, Prof. Aryan Mokhtari and Prof. Ali Jadbabaie. Prior to joining UCSB in January 2017, I recieved my Master's degree in EE from UCLA in 2016. I also got my Bachelor's degree in EE from Sharif University of Technology in 2014.

Research Interests

I am broadly interested in the area of large-scale distributed learning. In particular, I have been working on several areas of Distributed Machine Learning with a focus on addressing the essential challenges in Federated Learning and Distributed Computing frameworks. I have developed and employed a diverse set of tools and techniques from large-scale optimization, statistical learning theory, probability theory, and communication and coding theory. My work leverages the interplay between these areas and looks for novel solutions.

News

• December 2020: New preprint on Federated Learning:

  • Straggler-Resilient Federated Learning: Leveraging the Interplay Between Statistical Accuracy and System Heterogeneity [paper]

• September 2020: New paper accepted to NeurIPS 2020:

  • Robust Federated Learning: The Case of Affine Distribution Shifts [paper] [poster]

• July 2020: I presented our recent work at ICML 2020 Workshop on Federated Learning:

  • Robust Federated Learning: The Case of Affine Distribution Shifts [paper] [poster]

• May 2020: New paper accepted to IEEE Transactions on Information Theory.

  • Coded Computing for Distributed Graph Analytics [paper]

• May 2020: New paper accepted to ISIT 2020.

  • Hierarchical Coded Gradient Aggregation for Learning at the Edge

• February 2020: Graduation-day talk at ITA 2020, San Diego, CA. [poster]

• January 2020: Our work will appear in AISTATS 2020.

  • FedPAQ: A Communication-Efficient Federated Learning Method with Periodic Averaging and Quantization [paper]

• September 2019: We will present our work in NeurIPS 2019.

  • Robust and Communication-Efficient Collaborative Learning [paper] [poster]

• July 2019: Our paper got accepted to IEEE Transactions on Signal Processing (TSP).

  • An Exact Quantized Decentralized Gradient Descent Algorithm [paper]

• July 2019: We will present our work at 2019 IEEE International Symposium on Information Theory (ISIT), Paris, France.

  • Tree Gradeint Coding [paper]

• July 2019: I presented a poster on our work CodedReduce at North American School of Information Theory, Boston University, Boston, MA.

  • CodedReduce: A Fast and Robust Framework for Gradient Aggregation in Distributed Learning [paper][poster]

• June 2019: I presented a poster on our work CodedReduce at International Conference on Machine Learning (ICML) Workshop on Coding Theory For Large-scale Machine Learning, Long Beach, CA.

  • CodedReduce: A Fast and Robust Framework for Gradient Aggregation in Distributed Learning [paper][poster]

• May 2019: I presented a poster on our work at Learning for Dynamical and Control Systems (L4DC) Conference, MIT, Cambridge, MA.

  • Robust and Communication-Efficient Collaborative Learning [poster]

• April 2019: Qualcomm Innovation Fellowship finalist