ICML 2021 accepted papers on Federated learning and privacy
43 of the 1,184 papers accepted at ICML 2021 (3.6%) are on Federated learning and privacy, compared with 2.0% at ICML 2020. They fall into 4 topics.
See ICML 2021 in the interactive map
Submitting to ICML? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Differential privacy · 20 papers
- The Distributed Discrete Gaussian Mechanism for Federated Learning with Secure Aggregation (25 citations)
- Practical and Private (Deep) Learning Without Sampling or Shuffling (21 citations)
- Leveraging Public Data for Practical Private Query Release (18 citations)
- Differentially Private Query Release Through Adaptive Projection (16 citations)
- Large Scale Private Learning via Low-rank Reparametrization (14 citations)
- Differentially Private Bayesian Inference for Generalized Linear Models (9 citations)
- Lossless Compression of Efficient Private Local Randomizers (7 citations)
- Accuracy, Interpretability, and Differential Privacy via Explainable Boosting (6 citations)
- Differentially Private Aggregation in the Shuffle Model: Almost Central Accuracy in Almost a Single Message (6 citations)
- Differentially-Private Clustering of Easy Instances (4 citations)
- PAPRIKA: Private Online False Discovery Rate Control (4 citations)
- Differentially Private Sliced Wasserstein Distance (3 citations)
and 8 more in the interactive map.
Communication-efficient FL · 9 papers
- Clustered Sampling: Low-Variance and Improved Representativity for Clients Selection in Federated Learning (42 citations)
- Federated Composite Optimization (16 citations)
- Bias-Variance Reduced Local SGD for Less Heterogeneous Federated Learning (13 citations)
- Federated Learning under Arbitrary Communication Patterns (13 citations)
- Federated Deep AUC Maximization for Hetergeneous Data with a Constant Communication Complexity (7 citations)
- Communication-Efficient Distributed SVD via Local Power Iterations (6 citations)
- FL-NTK: A Neural Tangent Kernel-based Framework for Federated Learning Analysis (4 citations)
- Cross-Gradient Aggregation for Decentralized Learning from Non-IID Data (3 citations)
- Debiasing Model Updates for Improving Personalized Federated Training
Privacy attacks and secure learning · 7 papers
- CRFL: Certifiably Robust Federated Learning against Backdoor Attacks (47 citations)
- DeepReDuce: ReLU Reduction for Fast Private Inference (22 citations)
- Privacy-Preserving Feature Selection with Secure Multiparty Computation (18 citations)
- When Does Data Augmentation Help With Membership Inference Attacks? (15 citations)
- HEMET: A Homomorphic-Encryption-Friendly Privacy-Preserving Mobile Neural Network Architecture (12 citations)
- Gradient Disaggregation: Breaking Privacy in Federated Learning by Reconstructing the User Participant Matrix (10 citations)
- Matrix Sketching for Secure Collaborative Machine Learning (1 citation)
Personalized and heterogeneous FL · 7 papers
- Ditto: Fair and Robust Federated Learning Through Personalization (216 citations)
- Personalized Federated Learning using Hypernetworks (84 citations)
- Exploiting Shared Representations for Personalized Federated Learning (69 citations)
- Data-Free Knowledge Distillation for Heterogeneous Federated Learning (66 citations)
- Federated Continual Learning with Weighted Inter-client Transfer (63 citations)
- Heterogeneity for the Win: One-Shot Federated Clustering (21 citations)
- Federated Learning of User Verification Models Without Sharing Embeddings (2 citations)
