h1

h2

h3

h4

h5
h6

RLeap

From Data-based to Model-based AI: Representation Learning for Planning

Grant period2020-10-01 - 2025-09-30
Funding bodyEuropean Union
Call numberERC-2019-ADG
Grant number885107
IdentifierG:(EU-Grant)885107

Note: Two of the main research threads in AI revolve around the development of data-based learners capable of inferring behavior and functions from experience and data, and model-based solvers capable of tackling well-defined but intractable models like SAT, classical planning, and Bayesian networks. Learners, and in particular deep learners, have achieved considerable success but result in black boxes that do not have the flexibility, transparency, and generality of their model-based counterparts. Solvers, on the other hand, require models which are hard to build by hand. RLeap is aimed at achieving an integration of learners and solvers in the context of planning by addressing and solving the problem of learning first-order planning representations from raw perceptions alone without using any prior symbolic knowledge. The ability to construct first-order symbolic representations and using them for expressing, communicating, achieving, and recognizing goals is a main component of human intelligence and a fundamental, open research problem in AI. The success of RLeap requires the development of radically new ideas and methods that will build on those of a number of related areas that include planning, learning, knowledge representation, combinatorial optimization and SAT. The approach to be pursued is based on a clear separation between learning the symbolic representations themselves, that is cast as a combinatorial problem, and learning the interpretations of those representations, that is cast as a supervised learning problem from targets obtained from the first part. RLeap will address both problems, not just in the planning setting but in the generalized planning setting as well where plans are general strategies. The project can make a significant difference in how general, explainable, and trustworthy AI can be understood and achieved. The PI has made key contribution to the main themes of the project that make him uniquely qualified to carry it forward.
   

Recent Publications

All known publications ...
Download: BibTeX | EndNote XML,  Text | RIS | 

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Preprint  ;
LTLf Synthesis on First-Order Agent Programs in Nondeterministic Environments
39. AAAI Conference on Artificial Intelligence, AAAI 2025, Philadelphia, PAPhiladelphia, PA, USA, 25 Feb 2025 - 4 Mar 20252025-02-252025-03-04 14 Seiten () [10.48550/arXiv.2410.00726]  GO OpenAccess   Download fulltextFulltext by arXiv.org BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Preprint  ;  ;
Learning Lifted STRIPS Models from Action Traces Alone: A Simple, General, and Scalable Solution
35. International Conference on Automated Planning and Scheduling, ICAPS 2025, MelbourneMelbourne, Australia, 9 Nov 2025 - 14 Nov 20252025-11-092025-11-14 13 Seiten () [10.48550/arXiv.2411.14995]  GO OpenAccess   Download fulltextFulltext by arXiv.org BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Preprint  ;  ;
Learning More Expressive General Policies for Classical Planning Domains
39. AAAI Conference on Artificial Intelligence, AAAI 2025, Philadelphia, PAPhiladelphia, PA, USA, 25 Feb 2025 - 4 Mar 20252025-02-252025-03-04 10 Seiten () [10.48550/arXiv.2403.11734]  GO OpenAccess   Download fulltextFulltext by arXiv.org BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Abstract (Extended abstract)/Contribution to a conference proceedings  ;
Strategy Synthesis for First-Order Agent Programs over Finite Traces
35th Nordic Workshop on Programming Theory (NWPT) : (Extended abstracts) : Copenhagen, November 6 - 8 2024 / Michael Kirkedal Thomsen, Maja Hanne Kirkeby, and Fritz Henglein (eds.)
35. Nordic Workshop on Programming Theory, NWPT, CopenhagenCopenhagen, Denmark, 6 Nov 2024 - 8 Nov 20242024-11-062024-11-08 4 Seiten ()  GO   Download fulltextFulltext BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Preprint  ;
Learning Sketch Decompositions in Planning via Deep Reinforcement Learning
39. Annual AAAI Conference on Artificial Intelligence, AAAI 2025, Philadelphia, PAPhiladelphia, PA, USA, 27 Feb 2025 - 4 Mar 20252025-02-272025-03-04 35 Seiten () [10.48550/arXiv.2412.08574]  GO OpenAccess  Download fulltext Files  Download fulltextFulltext by arXiv.org BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Poster  ;  ;  ;
Equivalence-Based Abstractions for Learning General Policies
ICAPS 2024 Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning, PRL, Meeting location, ()  GO BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Contribution to a book/Contribution to a conference proceedings  ;  ;
Learning to Ground Existentially Quantified Goals
KR 2024 : proceedings of the 21st International Conference on Principles of Knowledge Representation and Reasoning : Hanoi, Vietnam. November 2-8, 2024 / Edited by Pierre Marquis, Magdalena Ortiz, Maurice Pagnucco
21. International Conference on Principles of Knowledge Representation and Reasoning, KR 2024, HanoiHanoi, Vietnam, 2 Nov 2024 - 8 Nov 20242024-11-022024-11-08 IJCAI Organization 856-866 () [10.24963/kr.2024/80]  GO BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Contribution to a book/Contribution to a conference proceedings  ;  ;  ;
Symmetries and Expressive Requirements for Learning General Policies
KR 2024 : proceedings of the 21st International Conference on Principles of Knowledge Representation and Reasoning : Hanoi, Vietnam. November 2-8, 2024 / Edited by Pierre Marquis, Magdalena Ortiz, Maurice Pagnucco
21. International Conference on Principles of Knowledge Representation and Reasoning, KR 2024, HanoiHanoi, Vietnam, 2 Nov 2024 - 8 Nov 20242024-11-022024-11-08 IJCAI Organization 845-855 () [10.24963/kr.2024/79]  GO BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Contribution to a book/Contribution to a conference proceedings  ;  ;  ;
Combined Task and Motion Planning via Sketch Decompositions
Proceedings of the Thirty-Fourth International Conference on Automated Planning and Scheduling ICAPS-24 : June 1–6, 2024, Alberta, Canada / Edited by Sara Bernardini, Christian Muise
34. International Conference on Automated Planning and Scheduling, ICAPS 2024, AlbertaAlberta, Canada, 1 Jun 2024 - 6 Jun 20242024-06-012024-06-06 Washington, DC : Association for the Advancement of Artificial Intelligence, Proceedings of the International Conference on Automated Planning and Scheduling 34, 123-132 () [10.1609/icaps.v34i1.31468]  GO BibTeX | EndNote: XML, Text | RIS

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Contribution to a book/Contribution to a conference proceedings  ;
Learning Generalized Policies for Fully Observable Non-Deterministic Planning Domains
Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence : Jeju, Korea, 3-9 August 2024 / edited by Kate Larson ; sponsored by International Joint Conferences on Artifical Intelligence (IJCAI)
33. International Joint Conference on Artificial Intelligence, IJCAI 2024, JejuJeju, South Korea, 3 Aug 2023 - 9 Aug 20232023-08-032023-08-09 Vienna, Austria : International Joint Conferences on Artificial Intelligence 6733-6742 () [10.24963/ijcai.2024/744]  GO BibTeX | EndNote: XML, Text | RIS

All known publications ...
Download: BibTeX | EndNote XML,  Text | RIS | 

 Record created 2021-10-10, last modified 2023-02-09
Similar records



Rate this document:

Rate this document:
1
2
3
4
5
 
(Not yet reviewed)