ICML 2025 accepted papers on Federated learning and privacy
106 of the 3,342 papers accepted at ICML 2025 (3.2%) are on Federated learning and privacy, compared with 3.8% at ICML 2024. They fall into 4 topics.
See ICML 2025 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Differential privacy · 41 papers
- Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-off (3 citations)
- Certification for Differentially Private Prediction in Gradient-Based Training (1 citation)
- Differentially Private Boxplots (1 citation)
- A Unified Theoretical Analysis of Private and Robust Offline Alignment: from RLHF to DPO
- Approximate Differential Privacy of the $\ell_2$ Mechanism
- Auditing $f$-differential privacy in one run
- Breaking the $n^{1.5}$ Additive Error Barrier for Private and Efficient Graph Sparsification via Private Expander Decomposition
- Cape: Context-Aware Prompt Perturbation Mechanism with Differential Privacy
- Certifiably Robust Model Evaluation in Federated Learning under Meta-Distributional Shifts
- Contrastive Private Data Synthesis via Weighted Multi-PLM Fusion
- Differential Privacy Guarantees of Markov Chain Monte Carlo Algorithms
- Differential Privacy Under Class Imbalance: Methods and Empirical Insights
and 29 more in the interactive map.
Personalized and heterogeneous FL · 34 papers
- Multi-Session Budget Optimization for Forward Auction-based Federated Learning (2 citations)
- Interaction-Aware Gaussian Weighting for Clustered Federated Learning (1 citation)
- One-Shot Heterogeneous Federated Learning with Local Model-Guided Diffusion Models (1 citation)
- Understanding the Statistical Accuracy-Communication Trade-off in Personalized Federated Learning with Minimax Guarantees (1 citation)
- BSemiFL: Semi-supervised Federated Learning via a Bayesian Approach
- FedBEns: One-Shot Federated Learning based on Bayesian Ensemble
- FedECADO: A Dynamical System Model of Federated Learning
- Federated Generalised Variational Inference: A Robust Probabilistic Federated Learning Framework
- Federated Incomplete Multi-view Clustering with Globally Fused Graph Guidance
- Federated Learning for Feature Generalization with Convex Constraints
- FOCoOp: Enhancing Out-of-Distribution Robustness in Federated Prompt Learning for Vision-Language Models
- Harnessing Heterogeneous Statistical Strength for Personalized Federated Learning via Hierarchical Bayesian Inference
and 22 more in the interactive map.
Communication-efficient FL · 16 papers
- Compressed and distributed least-squares regression: convergence rates with applications to federated learning (104 citations)
- Federated In-Context Learning: Iterative Refinement for Improved Answer Quality
- Ferret: Federated Full-Parameter Tuning at Scale for Large Language Models
- FSL-SAGE: Accelerating Federated Split Learning via Smashed Activation Gradient Estimation
- Generalization in Federated Learning: A Conditional Mutual Information Framework
- NTK-DFL: Enhancing Decentralized Federated Learning in Heterogeneous Settings via Neural Tangent Kernel
- Privacy-Preserving Federated Convex Optimization: Balancing Partial-Participation and Efficiency via Noise Cancellation
- Scaffold with Stochastic Gradients: New Analysis with Linear Speed-Up
- SecEmb: Sparsity-Aware Secure Federated Learning of On-Device Recommender System with Large Embedding
- The Panaceas for Improving Low-Rank Decomposition in Communication-Efficient Federated Learning
- Widening the Network Mitigates the Impact of Data Heterogeneity on FedAvg
- FedSMU: Communication-Efficient and Generalization-Enhanced Federated Learning through Symbolic Model Updates
and 4 more in the interactive map.
Privacy attacks and secure learning · 15 papers
- Textual Unlearning Gives a False Sense of Unlearning (1 citation)
- A Certified Unlearning Approach without Access to Source Data
- An Efficient Private GPT Never Autoregressively Decodes
- Certified Unlearning for Neural Networks
- EgoPrivacy: What Your First-Person Camera Says About You?
- PASS: Private Attributes Protection with Stochastic Data Substitution
- Synthesizing Privacy-Preserving Text Data via Finetuning *without* Finetuning Billion-Scale LLMs
- An Effective and Secure Federated Multi-View Clustering Method with Information-Theoretic Perspective
- EncryptedLLM: Privacy-Preserving Large Language Model Inference via GPU-Accelerated Fully Homomorphic Encryption
- Gradient Inversion of Multimodal Models
- Janus: Dual-Server Multi-Round Secure Aggregation with Verifiability for Federated Learning
- Kona: An Efficient Privacy-Preservation Framework for KNN Classification by Communication Optimization
and 3 more in the interactive map.