Unverified50% confidenceBenchmarkExact time
Among the approximately 2,000 papers examined, more than 1,100 had at least one key claim independently verified, of which 266 were fully reproduced, 632 were partially reproduced through small-scale experiments, and approximately 4,000 individual claims were confirmed.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/12/2026
First Seen
Valid until: 11/10/2026
Sources
AI智能体复现ICML 2026论文：34%被验证，23%存在问题
Related Claims
VerifiedIn 2014, the NeurIPS program committee conducted a review consistency experiment in which approximately 10% (about 166 papers) were reviewed by two independent program committees, and the two groups' acceptance decisions were only approximately 57% consistent.65% similarUnverifiedApproximately 23% of papers could not be fully reproduced as claimed, and at least 496 papers had at least one claim that was falsified or disputed.65% similarUnverifiedSerenity's methodology requires that when using research reports for analysis, more than 1,000 reports are recommended for major tracks, with 200 reports as the starting quantity.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/737191API
curl https://kongchang.com/api/v1/knowledge/claims/737191MCP
get_claim(id=737191)