GFlowNet-EM for Learning Compositional Latent Variable Models

Edward J Hu, Nikolay Malkin, Moksh Jain, Katie E Everett, Alexandros Graikos, Yoshua Bengio
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:13528-13549, 2023.

Abstract

Latent variable models (LVMs) with discrete compositional latents are an important but challenging setting due to a combinatorially large number of possible configurations of the latents. A key tradeoff in modeling the posteriors over latents is between expressivity and tractable optimization. For algorithms based on expectation-maximization (EM), the E-step is often intractable without restrictive approximations to the posterior. We propose the use of GFlowNets, algorithms for sampling from an unnormalized density by learning a stochastic policy for sequential construction of samples, for this intractable E-step. By training GFlowNets to sample from the posterior over latents, we take advantage of their strengths as amortized variational inference algorithms for complex distributions over discrete structures. Our approach, GFlowNet-EM, enables the training of expressive LVMs with discrete compositional latents, as shown by experiments on non-context-free grammar induction and on images using discrete variational autoencoders (VAEs) without conditional independence enforced in the encoder.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-hu23c, title = {{GF}low{N}et-{EM} for Learning Compositional Latent Variable Models}, author = {Hu, Edward J and Malkin, Nikolay and Jain, Moksh and Everett, Katie E and Graikos, Alexandros and Bengio, Yoshua}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {13528--13549}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/hu23c/hu23c.pdf}, url = {https://proceedings.mlr.press/v202/hu23c.html}, abstract = {Latent variable models (LVMs) with discrete compositional latents are an important but challenging setting due to a combinatorially large number of possible configurations of the latents. A key tradeoff in modeling the posteriors over latents is between expressivity and tractable optimization. For algorithms based on expectation-maximization (EM), the E-step is often intractable without restrictive approximations to the posterior. We propose the use of GFlowNets, algorithms for sampling from an unnormalized density by learning a stochastic policy for sequential construction of samples, for this intractable E-step. By training GFlowNets to sample from the posterior over latents, we take advantage of their strengths as amortized variational inference algorithms for complex distributions over discrete structures. Our approach, GFlowNet-EM, enables the training of expressive LVMs with discrete compositional latents, as shown by experiments on non-context-free grammar induction and on images using discrete variational autoencoders (VAEs) without conditional independence enforced in the encoder.} }
Endnote
%0 Conference Paper %T GFlowNet-EM for Learning Compositional Latent Variable Models %A Edward J Hu %A Nikolay Malkin %A Moksh Jain %A Katie E Everett %A Alexandros Graikos %A Yoshua Bengio %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-hu23c %I PMLR %P 13528--13549 %U https://proceedings.mlr.press/v202/hu23c.html %V 202 %X Latent variable models (LVMs) with discrete compositional latents are an important but challenging setting due to a combinatorially large number of possible configurations of the latents. A key tradeoff in modeling the posteriors over latents is between expressivity and tractable optimization. For algorithms based on expectation-maximization (EM), the E-step is often intractable without restrictive approximations to the posterior. We propose the use of GFlowNets, algorithms for sampling from an unnormalized density by learning a stochastic policy for sequential construction of samples, for this intractable E-step. By training GFlowNets to sample from the posterior over latents, we take advantage of their strengths as amortized variational inference algorithms for complex distributions over discrete structures. Our approach, GFlowNet-EM, enables the training of expressive LVMs with discrete compositional latents, as shown by experiments on non-context-free grammar induction and on images using discrete variational autoencoders (VAEs) without conditional independence enforced in the encoder.
APA
Hu, E.J., Malkin, N., Jain, M., Everett, K.E., Graikos, A. & Bengio, Y.. (2023). GFlowNet-EM for Learning Compositional Latent Variable Models. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:13528-13549 Available from https://proceedings.mlr.press/v202/hu23c.html.

Related Material