CVPR rules vs. frontier models: must authors compare with a model released after submission?
Michael_J_Black · x · 2026-10-30
Michael Black highlights a policy gap in AI conference reviewing: CVPR 2026 guidelines state that failing to cite or beat an arXiv paper that appeared after submission should not be the sole grounds for rejection. But what if a large commercial or open-source model released post-submission might solve the problem—must authors rerun comparisons, and does skipping Astra justify rejection? Current rules say no for academic models but are silent on frontier models. He calls for clarifying the policy before CVPR 2027 reviewing begins.
More from Safety
- Anthropic hacked to deliver remote access trojan to Claude developer laptops — bashtoni · 2026-10-30
- Binance breach hits ~$6.3B: compromised DevOps laptop, ~$3B frozen on-chain — banteg · 2026-10-30
- Austria blocks digital red-tape cut over notary revenue, mocking EU AI ambitions — mitsuhiko · 2026-10-30
- Claude Code config reveals block on recreating existing video games, sparking developer backlash — mark_k · 2026-10-30
- How do you sandbox your AI coding agent? Reddit thread compares isolation setups — Chooseyourmindset · 2026-10-30
- Polymarket supply chain attack losses tracked at $2.7M, all visitors exposed — banteg · 2026-10-30