SkillMimicGen: Automated Demonstration Generation for Efficient Skill Learning and Deployment

Caelan Reed Garrett, Ajay Mandlekar, Bowen Wen, Dieter Fox
Proceedings of The 8th Conference on Robot Learning, PMLR 270:2750-2790, 2025.

Abstract

Imitation learning from human demonstrations is an effective paradigm for robot manipulation, but acquiring large datasets is costly and resource-intensive, especially for long-horizon tasks. To address this issue, we propose SkillGen, an automated system for generating demonstration datasets from a few human demos. SkillGen segments human demos into manipulation skills, adapts these skills to new contexts, and stitches them together through free-space transit and transfer motion. We also propose a Hybrid Skill Policy (HSP) framework for learning skill initiation, control, and termination components from SkillGen datasets, enabling skills to be sequenced using motion planning at test-time. We demonstrate that SkillGen greatly improves data generation and policy learning performance over a state-of-the-art data generation framework, resulting in the capability produce data for large scene variations, including clutter, and agents that are on average 24% more successful. We demonstrate the efficacy of SkillGen by generating over 24K demonstrations across 18 task variants in simulation from just 60 human demonstrations, and training proficient, often near-perfect, HSP agents. Finally, we apply SkillGen to 3 real-world manipulation tasks and demonstrate zero-shot sim-to-real transfer on a long-horizon assembly task. Videos, and more at https://skillgen.github.io.

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-garrett25a, title = {SkillMimicGen: Automated Demonstration Generation for Efficient Skill Learning and Deployment}, author = {Garrett, Caelan Reed and Mandlekar, Ajay and Wen, Bowen and Fox, Dieter}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {2750--2790}, year = {2025}, editor = {Agrawal, Pulkit and Kroemer, Oliver and Burgard, Wolfram}, volume = {270}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v270/main/assets/garrett25a/garrett25a.pdf}, url = {https://proceedings.mlr.press/v270/garrett25a.html}, abstract = {Imitation learning from human demonstrations is an effective paradigm for robot manipulation, but acquiring large datasets is costly and resource-intensive, especially for long-horizon tasks. To address this issue, we propose SkillGen, an automated system for generating demonstration datasets from a few human demos. SkillGen segments human demos into manipulation skills, adapts these skills to new contexts, and stitches them together through free-space transit and transfer motion. We also propose a Hybrid Skill Policy (HSP) framework for learning skill initiation, control, and termination components from SkillGen datasets, enabling skills to be sequenced using motion planning at test-time. We demonstrate that SkillGen greatly improves data generation and policy learning performance over a state-of-the-art data generation framework, resulting in the capability produce data for large scene variations, including clutter, and agents that are on average 24% more successful. We demonstrate the efficacy of SkillGen by generating over 24K demonstrations across 18 task variants in simulation from just 60 human demonstrations, and training proficient, often near-perfect, HSP agents. Finally, we apply SkillGen to 3 real-world manipulation tasks and demonstrate zero-shot sim-to-real transfer on a long-horizon assembly task. Videos, and more at https://skillgen.github.io.} }
Endnote
%0 Conference Paper %T SkillMimicGen: Automated Demonstration Generation for Efficient Skill Learning and Deployment %A Caelan Reed Garrett %A Ajay Mandlekar %A Bowen Wen %A Dieter Fox %B Proceedings of The 8th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Pulkit Agrawal %E Oliver Kroemer %E Wolfram Burgard %F pmlr-v270-garrett25a %I PMLR %P 2750--2790 %U https://proceedings.mlr.press/v270/garrett25a.html %V 270 %X Imitation learning from human demonstrations is an effective paradigm for robot manipulation, but acquiring large datasets is costly and resource-intensive, especially for long-horizon tasks. To address this issue, we propose SkillGen, an automated system for generating demonstration datasets from a few human demos. SkillGen segments human demos into manipulation skills, adapts these skills to new contexts, and stitches them together through free-space transit and transfer motion. We also propose a Hybrid Skill Policy (HSP) framework for learning skill initiation, control, and termination components from SkillGen datasets, enabling skills to be sequenced using motion planning at test-time. We demonstrate that SkillGen greatly improves data generation and policy learning performance over a state-of-the-art data generation framework, resulting in the capability produce data for large scene variations, including clutter, and agents that are on average 24% more successful. We demonstrate the efficacy of SkillGen by generating over 24K demonstrations across 18 task variants in simulation from just 60 human demonstrations, and training proficient, often near-perfect, HSP agents. Finally, we apply SkillGen to 3 real-world manipulation tasks and demonstrate zero-shot sim-to-real transfer on a long-horizon assembly task. Videos, and more at https://skillgen.github.io.
APA
Garrett, C.R., Mandlekar, A., Wen, B. & Fox, D.. (2025). SkillMimicGen: Automated Demonstration Generation for Efficient Skill Learning and Deployment. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:2750-2790 Available from https://proceedings.mlr.press/v270/garrett25a.html.

Related Material