Unverified50% confidenceFactExact time
Recent research (such as DER and SR-SAC) has attempted to push UTD up to 4, 10, or even higher in order to squeeze out greater data efficiency.
1
Sources
50%
Confidence
Long-term
Relevance
9/26/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedReplay Ratio (UTD) refers to the number of gradient update steps the Learner performs per step of environment data collected; standard DQN typically uses a value of 1 or lower.65% similarUnverifiedAfter data deduplication (using locality-sensitive hashing algorithms such as MinHash and SimHash), model convergence speed can improve by 10–20%.61% similarUnverifiedAcross a total of 446 pairwise comparisons, CTD statistically significantly outperforms current state-of-the-art methods in 72.9% of cases.61% similarUnverifiedAs model scale increases, the accuracy loss from using DA shows a trend of diminishing.57% similarUnverifiedMeta、NVIDIA等公司的研究表明RAG在知识密集型任务上的准确率比纯参数记忆提高20%-40%57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/952690API
curl https://kongchang.com/api/v1/knowledge/claims/952690MCP
get_claim(id=952690)