Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified50% confidenceTradeoffExact time

多步回报使用n步真实回报替代单步bootstrap，n=1偏差高方差低，n→∞退化为蒙特卡洛回报，实践中n=3到5是较好的折中

1

Sources

50%

Confidence

Long-term

Relevance

8/20/2026

First Seen

Sources

Rainbow DQN没有新想法：值函数强化学习算法演进全解析

Related Claims

UnverifiedReLU激活函数导数恒为1或0，与残差连接ResNet提供的梯度短路通道，是解决梯度消失的两大关键创新64% similar UnverifiedReAct范式相比传统的计划-执行分离模式，最大优势在于容错性，可根据工具返回结果动态修正后续计划63% similar Unverified尾调用解释器的优势在于分布式的分支预测（每个指令处理函数有自己的跳转点）和避免栈增长（通过尾调用优化复用栈帧）61% similar Unverified将整数转换为浮点数、执行浮点除法、再转换回整数虽然增加两次类型转换开销，但在特定场景下整体吞吐量反而更优59% similar UnverifiedReLU是目前最流行的激活函数，将负值置零、正值不变，计算高效且有效缓解梯度消失59% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/778673

API

curl https://kongchang.com/api/v1/knowledge/claims/778673

MCP

get_claim(id=778673)