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% confidenceFactExact time

假设数据服从高斯分布，最大化似然等价于最小化均方误差

1

Sources

50%

Confidence

Long-term

Relevance

7/12/2026

First Seen

Sources

机器学习核心概念深度解析：那些让人豁然开朗的顿悟时刻

redditr/learnmachinelearning7/11/2026

Related Claims

Unverified假设数据服从伯努利或多项分布，最大化似然等价于最小化交叉熵损失函数80% similar Unverified均匀量化引入的误差近似服从均匀分布，其方差为Δ²/12（Δ为量化步长）71% similar Unverified当大量用户使用相似prompt调用同一底层模型时，因底层权重和概率分布一致，输出的趋同几乎是数学上的必然71% similar Unverified对于参数量有限的小模型，数据质量对最终性能的影响甚至超过数据量本身70% similar Unverified标准化提取比例数字越高（如10:1、20:1）代表浓缩倍数越大，但高浓缩可能损失协同性次要成分（entourage effect），全谱提取物（full-spectrum）活性更接近原植物但含量标注模糊，二者需权衡70% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/494151

API

curl https://kongchang.com/api/v1/knowledge/claims/494151

MCP

get_claim(id=494151)