# Moment Matching Deep Contrastive Latent Variable Models Source: [Moment Matching Deep Contrastive Latent Variable Models]() Collection index: [llms.txt]() Data updated: 2025-10-28T13:56:51.898Z Authors: Ethan Weinberger, Nicasia Beebe-Wang, Su-In Lee Published: 2022-02-21T22:27:30.000Z ## Abstract In the contrastive analysis (CA) setting, machine learning practitioners are specifically interested in discovering patterns that are enriched in a target dataset as compared to a background dataset generated from sources of variation irrelevant to the task at hand. For example, a biomedical data analyst may seek to understand variations in genomic data only present among patients with a given disease as opposed to those also present in healthy control subjects. Such scenarios have motivated the development of contrastive latent variable models to isolate variations unique to these target datasets from those shared across the target and background datasets, with current state of the art models based on the variational autoencoder (VAE) framework. However, previously proposed models do not explicitly enforce the constraints on latent variables underlying CA, potentially leading to the undesirable leakage of information between the two sets of latent variables. Here we propose the moment matching contrastive VAE (MM-cVAE), a reformulation of the VAE for CA that uses the maximum mean discrepancy to explicitly enforce two crucial latent variable constraints underlying CA. On three challenging CA tasks we find that our method outperforms the previous state-of-the-art both qualitatively and on a set of quantitative metrics. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Reconstruction on FFHQ (SSIM, LPIPS, ID-Sim)](): Image Reconstruction; FFHQ; SSIM: 37. - [Latent Factor Separation on FFHQ Head Pose: Frontal vs. Right/Left](): Latent Factor Separation; FFHQ Head Pose: Frontal vs. Right/Left; C Metric: 0.73. - [Latent Factor Separation on FFHQ Smile vs. Non-Smiling](): Latent Factor Separation; FFHQ Smile vs. Non-Smiling; C Score: 0.66. - [Latent Factor Separation on FFHQ No Glasses vs. Glasses](): Latent Factor Separation; FFHQ No Glasses vs. Glasses; C Score: 0.58. - [Latent Factor Separation on FFHQ (Male vs Female)](): Latent Factor Separation; FFHQ Male vs. Female; C Score: 0.64. - [Salient Attribute Swapping on FFHQ Glasses Attribute](): Salient Attribute Swapping; FFHQ Glasses Attribute; Accuracy: 11.8. - [Attribute Swapping on FFHQ Swap X to Y high-quality (test)](): Attribute Swapping; FFHQ Swap X to Y high-quality (test); ID Similarity: 0.085. - [Attribute Swapping on FFHQ Swap Y to X high-quality (test)](): Attribute Swapping; FFHQ Swap Y to X high-quality (test); ID-Sim: 0.077. - [Image Reconstruction on FFHQ X dataset high-quality (test)](): Image Reconstruction; FFHQ X dataset high-quality (test); LPIPS: 0.6. - [Image Reconstruction on FFHQ Y dataset high-quality (test)](): Image Reconstruction; FFHQ Y dataset high-quality (test); LPIPS: 0.608. ## Original sources and implementation [Original paper]() [PDF]()