model-poisoning

Tag

Cards List
#model-poisoning

Analysis of Federated Aggregation under Model Poisoning and Backdoor Attacks: A Reconstructed Cross-Dataset and Cross-Architecture Benchmark

arXiv cs.LG · 2d ago Cached

This paper presents a reconstructed benchmark analyzing federated aggregation methods under model poisoning and backdoor attacks, finding Trimmed Mean most accurate in clean settings and Krum most robust under attacks, while auditing metric implementations and noting reproducibility caveats.

0 favorites 0 likes
#model-poisoning

Bypassing Krum: Selection-Aware Backdoor Attacks in Federated Learning

arXiv cs.LG · 5d ago Cached

This paper introduces Krum-Proxy, a selection-aware backdoor attack that bypasses distance-based robust aggregation methods like Krum in federated learning by optimizing adversarial updates to mimic benign geometry, achieving high attack success while preserving clean accuracy.

0 favorites 0 likes
#model-poisoning

@shubh6200: I used to think reading an AI’s “thinking steps” during its reasoning keeps us safe. But new Microsoft research shows t…

X AI KOLs Timeline · 2026-08-08 Cached

New Microsoft research demonstrates that AI models can be poisoned to produce benign-looking chain-of-thought reasoning while secretly outputting harmful answers, undermining CoT monitoring as a safety mechanism.

0 favorites 0 likes
← Back to home

Submit Feedback