machine learning

变分信息瓶颈

Variational Information Bottleneck

A theoretical framework that seeks to find an optimal trade-off between the accuracy of a representation and the amount of information it preserves about the input variable, often formulated as a variational optimization problem. / 一种寻求表示的准确性与关于输入变量所保留信息量之间最优权衡的理论框架，通常被表述为变分优化问题。

Knowledge type: 数学理论
First proposed or discovered
Proposer: Turkey Farm member
Creator: 农场主
Last updated: 2026-10-30 17:06:20

Sub-knowledge points

Information bottleneck principleVariational inferenceMultimodal representation learning

Higher-order knowledge points

Deep learning theoryRepresentation learning

Main parameters and performance indicators

Input variable X: The source data to be compressed.Target variable Y: The variable to be predicted or preserved.Rate-R distortion trade-off: The balance between compression rate and representation accuracy.

Open problems

Theoretical analysis of the convergence properties of variational inference in the I-B framework.Scalability to high-dimensional multimodal data.

Related literature

ranked by community stars

Learning discriminative prototypes via Variational Information Bottleneck for multimodal cancer survival predictionOpen source

变分信息瓶颈

Sub-knowledge points

Higher-order knowledge points

Main parameters and performance indicators

Open problems

Related products and links