Disentangled representations: towards interpretation of sex determination from hip bone

Zou, Kaifeng; Faisan, Sylvain; Heitz, Fabrice; Epain, Marie; Croisille, Pierre; Fanton, Laurent; Valette, Sébastien

doi:10.1007/s00371-022-02755-0

Search citation statements

Order By: Relevance

Paper Sections

Select...

Introduction1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2023

Publication Types

Select...

Article1

Other1

Relationship

Self Cite1

Independent1

Authors

Journals

Cited by 2 publications

(1 citation statement)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is a key challenge to learn disentangled representations where high-level factors would be independently and explicitly encoded [1]. Disentangled representations allow us to manipulate data by modifying high level factors, thus paving the way to easier interpretation of the influence of these factors [2]. It has also been shown that these representations may be more sample-efficient, less sensitive to nuisance variables, and better in terms of generalization [3].…”

Section: Introductionmentioning

confidence: 99%

Disentangling high-level factors and their features with conditional vector quantized VAEs

Zou

Faisan

Heitz

et al. 2023

Pattern Recognition Letters

Self Cite

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Disentangling high-level factors and their features with conditional vector quantized VAEs

Zou

Faisan

Heitz

et al. 2023

Pattern Recognition Letters

Self Cite

View full text Add to dashboard Cite

Joint Disentanglement of Labels and Their Features with VAE

Zou

Faisan

Heitz

et al. 2022

2022 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Most of previous semi-supervised methods that seek to obtain disentangled representations using variational autoencoders divide the latent representation into two components: the non-interpretable part and the disentangled part that explicitly models the factors of interest. With such models, features associated with high-level factors are not explicitly modeled, and they can either be lost, or at best entangled in the other latent variables, thus leading to bad disentanglement properties. To address this problem, we propose a novel conditional dependency structure where both the labels and their features belong to the latent space. We show using the CelebA dataset that the proposed model can learn meaningful representations, and we provide quantitative and qualitative comparisons with other approaches that show the effectiveness of the proposed method.

show abstract

Disentangled representations: towards interpretation of sex determination from hip bone

Cited by 2 publications

References 34 publications

Disentangling high-level factors and their features with conditional vector quantized VAEs

Disentangling high-level factors and their features with conditional vector quantized VAEs

Joint Disentanglement of Labels and Their Features with VAE

Contact Info

Product

Resources

About