ICLR 2025 accepted papers on Federated learning and privacy
64 of the 3,703 papers accepted at ICLR 2025 (1.7%) are on Federated learning and privacy, compared with 2.7% at ICLR 2024. They fall into 4 topics.
See ICLR 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 · 24 papers
- PFGuard: A Generative Framework with Privacy and Fairness Safeguards (2 citations)
- Privacy-Preserving Personalized Federated Prompt Learning for Multimodal Large Language Models (2 citations)
- Differentially Private Steering for Large Language Model Alignment (1 citation)
- Optimality of Matrix Mechanism on $\ell_p^p$-metric (1 citation)
- The Last Iterate Advantage: Empirical Auditing and Principled Heuristic Analysis of Differentially Private SGD (1 citation)
- Tighter Privacy Auditing of DP-SGD in the Hidden State Threat Model (1 citation)
- Data-adaptive Differentially Private Prompt Synthesis for In-Context Learning
- Differentially Private Federated Learning with Time-Adaptive Privacy Spending
- Differentially private learners for heterogeneous treatment effects
- Differentially private optimization for non-decomposable objective functions
- DiSK: Differentially Private Optimizer with Simplified Kalman Filter for Noise Reduction
- Does Training with Synthetic Data Truly Protect Privacy?
and 12 more in the interactive map.
Personalized and heterogeneous FL · 19 papers
- Federated Domain Generalization with Data-free On-server Matching Gradient (1 citation)
- On the Byzantine-Resilience of Distillation-Based Federated Learning (1 citation)
- Selective Aggregation for Low-Rank Adaptation in Federated Learning (1 citation)
- Unlocking the Potential of Model Calibration in Federated Learning (1 citation)
- Connecting Federated ADMM to Bayes
- Decoupled Subgraph Federated Learning
- Enhancing Clustered Federated Learning: Integration of Strategies and Improved Methodologies
- Enhancing Federated Domain Adaptation with Multi-Domain Prototype-Based Federated Fine-Tuning
- Federated Class-Incremental Learning: A Hybrid Approach Using Latent Exemplars and Data-Free Techniques to Address Local and Global Forgetting
- Federated Residual Low-Rank Adaptation of Large Language Models
- FedTMOS: Efficient One-Shot Federated Learning with Tsetlin Machine
- GRAIN: Exact Graph Reconstruction from Gradients
and 7 more in the interactive map.
Communication-efficient FL · 15 papers
- Achieving Dimension-Free Communication in Federated Learning via Zeroth-Order Optimization
- Can Textual Gradient Work in Federated Learning?
- Debiasing Federated Learning with Correlated Client Participation
- Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees
- DICE: Data Influence Cascade in Decentralized Learning
- Event-Driven Online Vertical Federated Learning
- Federated $Q$-Learning with Reference-Advantage Decomposition: Almost Optimal Regret and Logarithmic Communication Cost
- Federated Granger Causality Learning For Interdependent Clients With State Space Representation
- LoCoDL: Communication-Efficient Distributed Learning with Local Training and Compression
- Methods with Local Steps and Random Reshuffling for Generally Smooth Non-Convex Federated Optimization
- On the Linear Speedup of Personalized Federated Reinforcement Learning with Shared Representations
- PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models
and 3 more in the interactive map.
Privacy attacks and secure learning · 6 papers
- MUSE: Machine Unlearning Six-Way Evaluation for Language Models (3 citations)
- CipherPrune: Efficient and Scalable Private Transformer Inference
- DCT-CryptoNets: Scaling Private Inference in the Frequency Domain
- DynFrs: An Efficient Framework for Machine Unlearning in Random Forest
- Bad-PFL: Exploiting Backdoor Attacks against Personalized Federated Learning
- Energy-based Backdoor Defense Against Federated Graph Learning