Unverified50% confidenceFactExact time
Estimating mutual information and conditional entropy under finite samples is extremely challenging: entropy estimation for high-dimensional continuous variables has systematic bias, and commonly used k-nearest-neighbor estimators (such as the Kraskov estimator) remain unstable when the hidden state dimension is high or serial correlation is strong.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe inherent uncertainty of the data distribution is measured by source entropy, which can be estimated with a finite-sample confidence lower bound using samples.77% similarVerifiedCross-entropy loss originates from KL divergence in information theory and measures the degree of difference between the true probability distribution and the predicted probability distribution.76% similarUnverifiedReflective open-ended questions have a high-entropy output space where the distribution of reasonable answers is dispersed, and intra-group noise masks cross-model signals; whereas low-entropy prompts have a concentrated set of reasonable answers, and models with different training distributions will produce quantifiable deviations.73% similarUnverifiedIn the paper, the amplification factor of model error for Q-value iteration under the risk-neutral setting is approximately 1/(1-γ), whereas under the recursive entropic risk setting, error amplification is more complex due to nonlinear exponential operations, resulting in an exponential dependence.72% similarUnverified标准交叉熵损失在目标占图像面积较小时会导致严重的类别不平衡问题，模型倾向于全部预测为背景70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1010266API
curl https://kongchang.com/api/v1/knowledge/claims/1010266MCP
get_claim(id=1010266)