Unverified50% confidenceFactExact time
Under the standard PAC framework, if the noise ratio exceeds a certain threshold, no algorithm can uniquely recover the target; list-decodable learning, however, allows outputting a short list containing the true parameters in polynomial time under high-noise conditions.
1
Sources
50%
Confidence
Long-term
Relevance
10/26/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder the setting of an exact verifier and a fixed source, moving all calls to before the first verification preserves all hard upper bounds on the number of calls; interleaved execution cannot break the hard upper bounds, although it may save infinitely many calls in expectation.66% similarUnverifiedThere is a counterintuitive conclusion: under the setting of an exact verifier and a fixed source, interleaved execution cannot break the hard upper bounds; but under a noisy verifier, the source-first protocol may lose an unbounded factor, indicating that curriculum-style intuition is unsafe in the noisy verification setting.65% similarUnverifiedThe fix for such problems cannot rely on 'making queries faster' or 'increasing the number of retries', but must introduce an explicit constraint on the time dimension at the model level, turning uncertain timing into a decidable state.64% similarUnverifiedIf no threshold can achieve the target precision at a reasonable recall, the problem is not the threshold—a re-ranker or better chunking is needed instead.64% similarUnverifiedMFA's alignment algorithm tends to drift when faced with long silences, and may incorrectly map silences to actual words when silence weights are set improperly.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1098118API
curl https://kongchang.com/api/v1/knowledge/claims/1098118MCP
get_claim(id=1098118)