7
[2310.00239] AdaptNet: Policy Adaptation for Physics-Based Character Control
source link: https://arxiv.org/abs/2310.00239
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
Computer Science > Graphics
[Submitted on 30 Sep 2023 (v1), last revised 15 Nov 2023 (this version, v3)]
AdaptNet: Policy Adaptation for Physics-Based Character Control
Motivated by humans' ability to adapt skills in the learning of new ones, this paper presents AdaptNet, an approach for modifying the latent space of existing policies to allow new behaviors to be quickly learned from like tasks in comparison to learning from scratch. Building on top of a given reinforcement learning controller, AdaptNet uses a two-tier hierarchy that augments the original state embedding to support modest changes in a behavior and further modifies the policy network layers to make more substantive changes. The technique is shown to be effective for adapting existing physics-based controllers to a wide range of new styles for locomotion, new task targets, changes in character morphology and extensive changes in environment. Furthermore, it exhibits significant increase in learning efficiency, as indicated by greatly reduced training times when compared to training from scratch or using other approaches that modify existing policies. Code is available at this https URL.
| Comments: | SIGGRAPH Asia 2023. Video: this https URL. Website: this https URL, this https URL |
| Subjects: | Graphics (cs.GR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2310.00239 [cs.GR] |
| (or arXiv:2310.00239v3 [cs.GR] for this version) | |
| https://doi.org/10.48550/arXiv.2310.00239 | |
| Journal reference: | ACM Transactions on Graphics 42, 6, Article 112.1522 (December 2023) |
| Related DOI: |
Submission history
From: Pei Xu [view email][v1] Sat, 30 Sep 2023 03:19:51 UTC (44,622 KB)
[v2] Mon, 9 Oct 2023 15:23:38 UTC (44,622 KB)
[v3] Wed, 15 Nov 2023 03:44:44 UTC (44,604 KB)
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK