DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 박대형 | - |
dc.contributor.author | Jang, Jaehwi | - |
dc.contributor.author | 장재휘 | - |
dc.date.accessioned | 2024-07-30T19:30:43Z | - |
dc.date.available | 2024-07-30T19:30:43Z | - |
dc.date.issued | 2024 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1096083&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/321378 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 김재철AI대학원, 2024.2,[v, 34p :] | - |
dc.description.abstract | We present the problem of inverse constraint learning (ICL), which recovers constraints from demonstrations to autonomously reproduce constrained skills in new scenarios. However, ICL suffers from an ill-posed nature, leading to inaccurate inference of constraints from demonstrations. To figure it out, we introduce a transferable constraint learning (TCL) algorithm that jointly infers a task-oriented reward and a task-agnostic constraint, enabling the generalization of learned skills. Our method TCL additively decomposes the overall reward recovered from an inverse reinforcement learning into a task reward and its residual as soft constraints, minimizing policy divergence between task-oriented policies and the demonstration to obtain a transferable constraint. Evaluating our method and five baselines in three simulated environments, we show TCL outperforms state-of-the-art IRL and ICL algorithms, achieving up to a 72% higher task-success rates with accurate decomposition compared to the next best approach in novel scenarios. Further, we demonstrate the robustness of TCL on two real-world robotic tasks. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 시연 학습▼a역제약 조건 학습▼a제약된 동작 계획법 | - |
dc.subject | Learning from demonstration▼aInverse constraint learning▼aConstrained motion planning | - |
dc.title | Inverse constraint learning and generalization by transferable reward decomposition | - |
dc.title.alternative | 전이 가능한 보상 분해를 통한 역제약 조건 학습 및 일반화 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :김재철AI대학원, | - |
dc.contributor.alternativeauthor | Park, Daehyung | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.