Vasisht Duddu
Ph.D. Student, Computer Science
Secure Systems Group
Cheriton School of Computer Science
CrySP Lab, University of Waterloo
vasisht.duddu at uwaterloo.ca



Home Publications MiscellaneousCV

Publications

2024

• Machine Learning Property Attestations using TEEs Paper DOI Code
  Vasisht Duddu, Oskari Järvinen, Lachlan J. Gunn, N. Asokan
  Poster @ IEEE Symposium on Security and Privacy (S&P), 2024
  Full paper in progress
  Keywords ⇒ Trusted Hardware Regulatory Compliance Machine Learning


• Espresso: Robust Concept Filtering in Text-to-Image Models Paper DOI Code
  Anudeep Das, Vasisht Duddu, Rui Zhang, N. Asokan
  Under submission
  Keywords ⇒ Concept Removal Adversarial Robustness Diffusion Models


• Attesting Distributional Properties of Training Data for Machine Learning Paper DOI Code
  Vasisht Duddu, Anudeep Das, Nora Khayata, Hossein Yalame, Thomas Schneider, N. Asokan
  European Symposium on Research in Computer Security (ESORICS), 2024
  Keywords ⇒ Regulatory Compliance Distribution Inference Cryptographic Attestation Machine Learning


• SoK: Unintended Interactions among Machine Learning Defenses and Risks Paper DOI Code
  Vasisht Duddu, Sebastian Szyller, N. Asokan
  IEEE Symposium on Security and Privacy (S&P), 2024
  Keywords ⇒ Overfitting Memorization Trustworthy Machine Learning Systematization


• GrOVe: Ownership Verification of Graph Neural Networks using Embeddings Paper DOI Code
  Asim Waheed, Vasisht Duddu, N. Asokan
  IEEE Symposium on Security and Privacy (S&P), 2024
  Keywords ⇒ Fingerprinting Ownership Verification Model Extraction Attacks Graph Neural Networks

2023

• Comprehension from Chaos: What Users Understand and Expect from Private Computation Paper DOI Code
  Bailey Kacsmar, Vasisht Duddu, Kyle Tilbury, Blase Ur, Florian Kerschbaum
  ACM Conference on Computer and Communications Security (CCS), 2023
  Keywords ⇒ User Study Private Computation Inference Attacks

2022

• On the Alignment of Group Fairness with Attribute Privacy Paper Code
  Jan Aalmoes^*, Vasisht Duddu^*, Antoine Boutet
  Under Submission
  Keywords ⇒ Attribute Privacy Inference Attacks Algorithmic Fairness Machine Learning


• Inferring Sensitive Attributes from Model Explanations Paper DOI Code
  Vasisht Duddu, Antoine Boutet
  ACM International Conference on Information and Knowledge Management (CIKM), 2022
  Keywords ⇒ Attribute Privacy Inference Attacks Explainable Machine Learning Deep Learning


• Good Artists Copy, Great Artists Steal: Model Extraction Attacks Against Image Translation GANs Paper Code
  Sebastian Szyller, Vasisht Duddu, Tommi Gröndahl, N. Asokan
  Technical Report 2021
  Keywords ⇒ Model Extraction Attack Image Translation Generative Adversarial Networks

2020

• Quantifying Privacy Leakage in Graph Embedding Paper DOI Code
  Vasisht Duddu, Antoine Boutet, Virat Shejwalkar
  EAI International Conference on Mobile and Ubiquitous Systems (MobiQuitous), 2020
  Workshop on Privacy Preserving Machine Learning - PriML and PPML Joint Edition. NeurIPS 2020.
  Keywords ⇒ Membership Privacy Attribute Privacy Graph Reconstruction Graph Machine Learning


• Fault Tolerance of Neural Networks in Adversarial Settings Paper DOI Code
  Vasisht Duddu, N. Rajesh Pillai, D. Vijay Rao, Valentina E. Balas
  Journal of Intelligent & Fuzzy Systems, 2020
  Keywords ⇒ Fault Tolerance Differential Privacy Adversarial Robustness Deep Learning


• Towards Enhancing Fault Tolerance in Neural Networks Paper DOI Code
  Vasisht Duddu, D. Vijay Rao, Valentina E. Balas
  EAI International Conference on Mobile and Ubiquitous Systems (MobiQuitous), 2020
  Keywords ⇒ Fault Tolerance Deep Learning

2018

• Stealing Neural Networks via Timing Side Channels Paper
  Vasisht Duddu, Debasis Samanta, D Vijay Rao, Valentina E. Balas
  Technical Report 2018 (Accepted for presentation in AI Village@DEFCON 27)
  Keywords ⇒ Model Extraction Attack Side Channel Leakage Deep Learning

Theses

• Towards Effective Measurement of Membership Privacy Risk for Machine Learning Models Thesis
  Vasisht Duddu
  Master's Thesis, University of Waterloo, 2022
  Technical Report: "SHAPr: An Efficient and Versatile Membership Privacy Risk Metric for Machine Learning" Report
  Vasisht Duddu, Sebastian Szyller, N. Asokan
  Keywords ⇒ Data Privacy Membership Privacy Inference Attacks Machine Learning