Fine-Tuning Generative Models as an Inference Method for Robotic Tasks

Orr Krupnik, Elisei Shafer, Tom Jurgenson, Aviv Tamar
Proceedings of The 7th Conference on Robot Learning, PMLR 229:866-886, 2023.

Abstract

Adaptable models could greatly benefit robotic agents operating in the real world, allowing them to deal with novel and varying conditions. While approaches such as Bayesian inference are well-studied frameworks for adapting models to evidence, we build on recent advances in deep generative models which have greatly affected many areas of robotics. Harnessing modern GPU acceleration, we investigate how to quickly adapt the sample generation of neural network models to observations in robotic tasks. We propose a simple and general method that is applicable to various deep generative models and robotic environments. The key idea is to quickly fine-tune the model by fitting it to generated samples matching the observed evidence, using the cross-entropy method. We show that our method can be applied to both autoregressive models and variational autoencoders, and demonstrate its usability in object shape inference from grasping, inverse kinematics calculation, and point cloud completion.

Cite this Paper

BibTeX
@InProceedings{pmlr-v229-krupnik23a, title = {Fine-Tuning Generative Models as an Inference Method for Robotic Tasks}, author = {Krupnik, Orr and Shafer, Elisei and Jurgenson, Tom and Tamar, Aviv}, booktitle = {Proceedings of The 7th Conference on Robot Learning}, pages = {866--886}, year = {2023}, editor = {Tan, Jie and Toussaint, Marc and Darvish, Kourosh}, volume = {229}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v229/krupnik23a/krupnik23a.pdf}, url = {https://proceedings.mlr.press/v229/krupnik23a.html}, abstract = {Adaptable models could greatly benefit robotic agents operating in the real world, allowing them to deal with novel and varying conditions. While approaches such as Bayesian inference are well-studied frameworks for adapting models to evidence, we build on recent advances in deep generative models which have greatly affected many areas of robotics. Harnessing modern GPU acceleration, we investigate how to quickly adapt the sample generation of neural network models to observations in robotic tasks. We propose a simple and general method that is applicable to various deep generative models and robotic environments. The key idea is to quickly fine-tune the model by fitting it to generated samples matching the observed evidence, using the cross-entropy method. We show that our method can be applied to both autoregressive models and variational autoencoders, and demonstrate its usability in object shape inference from grasping, inverse kinematics calculation, and point cloud completion.} }
Endnote
%0 Conference Paper %T Fine-Tuning Generative Models as an Inference Method for Robotic Tasks %A Orr Krupnik %A Elisei Shafer %A Tom Jurgenson %A Aviv Tamar %B Proceedings of The 7th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2023 %E Jie Tan %E Marc Toussaint %E Kourosh Darvish %F pmlr-v229-krupnik23a %I PMLR %P 866--886 %U https://proceedings.mlr.press/v229/krupnik23a.html %V 229 %X Adaptable models could greatly benefit robotic agents operating in the real world, allowing them to deal with novel and varying conditions. While approaches such as Bayesian inference are well-studied frameworks for adapting models to evidence, we build on recent advances in deep generative models which have greatly affected many areas of robotics. Harnessing modern GPU acceleration, we investigate how to quickly adapt the sample generation of neural network models to observations in robotic tasks. We propose a simple and general method that is applicable to various deep generative models and robotic environments. The key idea is to quickly fine-tune the model by fitting it to generated samples matching the observed evidence, using the cross-entropy method. We show that our method can be applied to both autoregressive models and variational autoencoders, and demonstrate its usability in object shape inference from grasping, inverse kinematics calculation, and point cloud completion.
APA
Krupnik, O., Shafer, E., Jurgenson, T. & Tamar, A.. (2023). Fine-Tuning Generative Models as an Inference Method for Robotic Tasks. Proceedings of The 7th Conference on Robot Learning, in Proceedings of Machine Learning Research 229:866-886 Available from https://proceedings.mlr.press/v229/krupnik23a.html.

Related Material