Education and Early Career

Dr Ali Shahin Shamsabadi pursued his doctoral studies under the supervision of Aurélien Bellet, Andrea Cavallaro, Adria Gascon, Hamed Haddadi, Matt Kusner, and Emmanuel Vincent. His PhD research focused on privacy-preserving machine learning and data security.

Professional Experience

After completing his PhD, Shamsabadi held positions as a Postdoctoral Fellow at the Vector Institute, supervised by Nicolas Papernot, and as a Research Scientist at The Alan Turing Institute, where he worked in the Safe and Ethical AI group under Adrian Weller. In these roles, he contributed to advancing privacy-preserving techniques in machine learning and artificial intelligence.

Currently, Shamsabadi is a Senior Research Scientist at Brave Software, focusing on enhancing user privacy and security in web technologies.

Research Contributions

Shamsabadi's research encompasses various aspects of data privacy, machine learning, and computer security. Notable publications include:

• –

  "A Hybrid Deep Learning Architecture for Privacy-Preserving Mobile Analytics": This paper presents a framework combining deep learning and privacy-preserving techniques for mobile data analytics. IEEE Internet of Things Journal, 2020

• –

  "QUOTIENT: Two-Party Secure Neural Network Training and Prediction": This work introduces a protocol for secure two-party computation in neural network training and inference. ACM Conference on Computer and Communications Security, 2019

• –

  "DarkneTZ: Towards Model Privacy at the Edge using Trusted Execution Environments": This research explores the use of trusted execution environments to protect machine learning models on edge devices. ACM International Conference on Mobile Systems, Applications, and Services, 2020

• –

  "ColorFool: Semantic Adversarial Colorization": This paper discusses adversarial attacks that manipulate image colorization to deceive machine learning models. Conference on Computer Vision and Pattern Recognition, 2020

• –

  "When the Curious Abandon Honesty: Federated Learning Is Not Private": This study examines privacy vulnerabilities in federated learning systems. IEEE European Symposium on Security and Privacy, 2023

Conference Presentations

Shamsabadi has presented his research at various esteemed conferences. At the USENIX Security Symposium 2023, he delivered talks on:

• –

  "GAP: Differentially Private Graph Neural Networks with Aggregation Perturbation": This presentation introduced a method for ensuring differential privacy in graph neural networks. USENIX Security Symposium 2023

• –

  "Tubes Among Us: Analog Attack on Automatic Speaker Identification": This talk discussed vulnerabilities in automatic speaker identification systems and proposed analog attack methods. USENIX Security Symposium 2023

Professional Recognition

Shamsabadi's contributions to the field have been recognized in various academic and professional circles. His work has been cited in numerous scholarly articles, reflecting his influence in the domains of data privacy and machine learning.

Online Presence

For more information on his research and publications, visit his personal website: https://alishahin.github.io/.