CuraSec

tag: Federated-Learning · 4 items

Learn FISGuard: Defense Against Membership Inference in Federated LLMs
2026-08-31 · arXiv cs.CR · source ↗ #federated-learning#privacy#llm-security
  • Engineer — Learn: Novel defense technique for federated fine-tuning pipelines; relevant if you run distributed LLM training with sensitive data, but no patch or configuration action needed today — research-stage only.
  • SOC/IR — Skip
  • Leader — Skip
Learn Federated Learning Watermarking + Lattice-Based Secure Aggregation Framework
2026-08-24 · arXiv cs.CR · source ↗ #federated-learning#watermarking#post-quantum
  • Engineer — Learn: Academic research proposing a combined data-provenance watermarking and post-quantum secure aggregation scheme for federated learning; no exploitation signals or patch action required, but relevant if you are designing or hardening an FL pipeline.
  • SOC/IR — Skip
  • Leader — Skip
Learn archived FLINT: 5G PHY-layer side channels fingerprint federated learning models
2026-07-20 · arXiv cs.CR · source ↗ #federated-learning#5g#side-channel
  • Engineer — Learn: Novel finding that 5G PDCCH scheduling metadata leaks enough temporal pattern to identify FL model architecture families, enabling targeted downstream attacks. No patch exists; worth factoring into FL-over-cellular deployment design (e.g., traffic shaping, scheduling obfuscation) before adopting this stack.
  • SOC/IR — Skip
  • Leader — Skip
Learn archived Federated Learning Privacy via Homomorphic Encryption & Differential Privacy
2026-07-13 · arXiv cs.CR · source ↗ #federated-learning#privacy-preserving-ml#homomorphic-encryption
  • Engineer — Learn: Academic architecture study combining homomorphic encryption and differential privacy for FL systems; no vulnerabilities or patches, but relevant for engineers designing privacy-preserving ML pipelines in healthcare or finance contexts.
  • SOC/IR — Skip
  • Leader — Learn: Research validates that FL with strong privacy controls can meet accuracy requirements in sensitive domains; useful background for evaluating AI/ML vendor privacy claims or shaping internal AI data-handling policy.