Research map / Federated learning and privacy / Privacy attacks and secure learning
FedRec++: Lossless Federated Recommendation with Explicit Feedback
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · ML100K | MAE: 0.7447± 0.0043 -> 0.7416± 0.0049 | Helped |
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · ML1M | MAE: 0.7263± 0.0010 -> 0.7196± 0.0013 | Helped |
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · NF5K5K | MAE: 0.7160± 0.0007 -> 0.7138± 0.0005 | Helped |
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · ML100K | RMSE: 0.9431± 0.0054 -> 0.9421± 0.0064 | Helped |
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · ML1M | RMSE: 0.9178± 0.0010 -> 0.9109± 0.0015 | Helped |
| Add denoising clients to remove gradients from sampled unrated items | FedRec with PMF · NF5K5K | RMSE: 0.9110± 0.0005 -> 0.9090± 0.0004 | Helped |
| Increase the sampled-unrated-item ratio ρ | FedRec · ML1M | MAE: 0.7217± 0.0011; 0.7239± 0.0011; 0.7263± 0.0010 | Hurt |
| Increase the sampled-unrated-item ratio ρ | FedRec · NF5K5K | MAE: 0.7148± 0.0004; 0.7152± 0.0005; 0.7160± 0.0007 | Hurt |
| Increase the sampled-unrated-item ratio ρ | FedRec · ML1M | RMSE: 0.9129± 0.0012; 0.9152± 0.0011; 0.9178± 0.0010 | Hurt |
| Increase the sampled-unrated-item ratio ρ | FedRec · NF5K5K | RMSE: 0.9102± 0.0005; 0.9104± 0.0005; 0.9110± 0.0005 | Hurt |
| Use 60% of clients instead of all clients during training | FedRec++ · ML100K and ML1M | MAE | No clear effect |
| Increase the number of denoising clients | FedRec++ · ML100K, ML1M, and NF5K5K | communication cost per client: 180; 286; 2073 | Hurt |
Closest papers by meaning
- Federated Recommendation with Additive Personalization (ICLR 2024 · similarity 0.89)
- SecEmb: Sparsity-Aware Secure Federated Learning of On-Device Recommender System with Large Embedding (ICML 2025 · similarity 0.88)
- Personalized Federated Collaborative Filtering: A Variational AutoEncoder Approach (AAAI 2025 · similarity 0.88)
- Federated Recommendation with Explicitly Encoding Item Bias (AAAI 2025 · similarity 0.87)
- Rethinking Byzantine Robustness in Federated Recommendation from Sparse Aggregation Perspective (AAAI 2025 · similarity 0.86)
- Multifaceted User Modeling in Recommendation: A Federated Foundation Models Approach (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.