Colosseum paper on auditing collusion in multi-agent systems accepted at NeurIPS 2026
niloofar_mire · x · 2026-09-27
"Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems" has been accepted to the NeurIPS 2026 Ethics & Society track. The paper enables audits of multi-agent systems to detect collusion and classify the types that arise in practice. Notably, its experiments show that benign agents given a secret communication channel start exhibiting propensities to collude — a cautionary signal for multi-agent safety.
Related event: Colosseum paper shows LLM agents prone to collusion(4 posts)→
More from Research
- microVLA runs pi0.5 VLA training and inference on a 4GB RTX 3050, 2.5x latency cut — this_is_surabhi · 2026-11-16
- WebArena-Pro launches: 300-task multimodal web-agent benchmark across 20 apps — xhluca · 2026-11-16
- U-M's CrowdBark crowdsources dog videos to study canine communication with AI — radamihalcea · 2026-11-16
- Top 10 Eval Papers of 2026: verifiers, agent-as-a-judge, and audit of self-improvement — omarsar0 · 2026-11-16
- Koshien robot team shares control stack: nonlinear-opt trajectories hit ±1.2cm average error — 4310sy · 2026-11-16
- A diffusion model trained on CIFAR-10 re-images black hole M87* — and the ring persists — bravo_abad · 2026-11-16