Papers (Please sign up for presentation here)

Privacy Attacks against Machine Learning

  1. Reconstructing Training Data with Informed Adversaries
  2. Extracting Training Data from Large Language Models
  3. Property Inference Attacks on Fully Connected Neural Networks using Permutation Invariant Representations
  4. Counterfactual Memorization in Neural Language Models
  5. Formalizing and Estimating Distribution Inference Risks
  6. Enhanced Membership Inference Attacks against Machine Learning Models
  7. Mitigating Membership Inference Attacks by Self-Distillation Through a Novel Ensemble Architecture
  8. Is Private Learning Possible with Instance Encoding?
  9. Submix: Practical Private Prediction For Large-scale Language Models
  10. Composition Attacks and Auxiliary Information in Data Privacy
  11. StolenEncoder: Stealing Pre-trained Encoders

Machine Learning Security (Instead of Privacy)

  1. Local Model Poisoning Attacks to Byzantine-Robust Federated Learning
  2. Spinning Language Models for Propaganda-As-A-Service
  3. Blind Backdoors in Deep Learning Models

Differential Privacy Theory

  1. Renyi Differential Privacy
  2. Numerical Composition of Differential Privacy
  3. Differentially Private Combinatorial Optimization
  4. Iterative Constructions and Private Data Release
  5. On the Rényi Differential Privacy of the Shuffle Model

Differential Privacy for Machine Learning

  1. Scalable Private Learning With PATE
  2. DPNAS: Neural Architecture Search for Deep Learning with Differential Privacy
  3. Hyperparameter Tuning with Renyi Differential Privacy
  4. Benchmarking Differential Privacy and Federated Learning for BERT Models
  5. Public Data-Assisted Mirror Descent for Private Model Training
  6. Adversary Instantiation: Lower Bounds for Differentially Private Machine Learning
  7. The Role of Adaptive Optimizers for Honest Private Hyperparameter Selection
  8. Differentially private fine-tuning of language models
  9. Large language models can be strong differentially private learners
  10. Large Scale Private Learning via Low-rank Reparametrization

Differential Privacy and Cryptography

  1. Strengthening Order Preserving Encryption with Differential Privacy
  2. Shrinkwrap: Efficient SQL Query Processing in Differentially Private Data Federations
  3. Differentially Private Oblivious RAM

Privacy and Systems

  1. Veil: Private Browsing Semantics Without Browser-side Assistance
  2. εpsolute: Efficiently Querying Databases While Providing Differential Privacy
  3. Packet scheduling with optional client privacy
  4. Data Privacy in Trigger-Action Systems
  5. εKTELO A Framework for Defining Differentially Private Computations
  6. PrivateSQL: a differentially private SQL query engine

Other General Privacy

  1. Privacy Engineering Meets Software Engineering. On the Challenges of Engineering Privacy By Design
  2. Towards formalizing the GDPR’s notion of singling out
  3. Differential privacy: An economic method for choosing epsilon
  4. Privacy Implications of Shuffling
  5. Causally Constrained Data Synthesis for Private Data Release
  6. DP-Sync: Hiding Update Patterns in Secure Outsourced Databases with Differential Privacy
  7. Kamino: Constraint-Aware Differentially Private Data Synthesis