给对比学习加不确定性估计，让模型知道自己的信心程度。
Probabilistic Variational Contrastive Learning
- 用变分推断重构对比学习，把嵌入变成可采样的概率分布。
- 在多个数据集上准确率不降反升，还解决了维度坍缩问题。
- 适合需要可信度评估的场景，如医疗图像识别、安全关键系统。
确定性对比学习方法（如 SimCLR、SupCon）虽性能顶尖，但缺乏不确定性量化机制。本文提出变分对比学习（VCL），一种无解码器框架，通过将 InfoNCE 损失视为替代重建项，并在单位超球面上添加均匀先验的 KL 正则项来最大化证据下界（ELBO）。我们采用投影正态分布建模近似后验 q_θ(z|x)，实现概率嵌入采样。两个实例化方法 VSimCLR 与 VSupCon 将确定性嵌入替换为 q_θ(z|x) 的采样结果，并引入归一化后的 KL 项到损失函数中。多基准测试表明，VCL 有效缓解维度坍缩，提升与类别标签的互信息，在分类准确率上匹配或超越确定性基线，同时通过后验模型提供有意义的不确定性估计。VCL 为对比学习提供了概率基础，成为新一类对比方法的基石。
原文摘要 · Abstract (English)
Deterministic embeddings learned by contrastive learning (CL) methods such as SimCLR and SupCon achieve state-of-the-art performance but lack a principled mechanism for uncertainty quantification. We propose Variational Contrastive Learning (VCL), a decoder-free framework that maximizes the evidence lower bound (ELBO) by interpreting the InfoNCE loss as a surrogate reconstruction term and adding a KL divergence regularizer to a uniform prior on the unit hypersphere. We model the approximate posterior $q_θ(z|x)$ as a projected normal distribution, enabling the sampling of probabilistic embeddings. Our two instantiation--VSimCLR and VSupCon--replace deterministic embeddings with samples from $q_θ(z|x)$ and incorporate a normalized KL term into the loss. Experiments on multiple benchmarks demonstrate that VCL mitigates dimensional collapse, enhances mutual information with class labels, and matches or outperforms deterministic baselines in classification accuracy, all the while providing meaningful uncertainty estimates through the posterior model. VCL thus equips contrastive learning with a probabilistic foundation, serving as a new basis for contrastive approaches.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。