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

Informazioni relative al progetto

RLeap

ID dell’accordo di sovvenzione: 885107

DOI

10.3030/885107

Progetto chiuso

Data della firma CE 24 Aprile 2020

Data di avvio 1 Ottobre 2020

Data di completamento 31 Marzo 2026

Finanziato da

• EXCELLENT SCIENCE - European Research Council (ERC)

Costo totale

€ 2 498 625,00

Contributo UE

€ 2 498 625,00

Coordinato da

RHEINISCH-WESTFAELISCHE TECHNISCHE HOCHSCHULE AACHEN
Germany

CORDIS fornisce collegamenti ai risultati finali pubblici e alle pubblicazioni dei progetti ORIZZONTE.

I link ai risultati e alle pubblicazioni dei progetti del 7° PQ, così come i link ad alcuni tipi di risultati specifici come dataset e software, sono recuperati dinamicamente da .OpenAIRE .

Risultati finali

Open Research Data Pilot (1)

Data Management Plan (DMP).

The Open Research Data Pilot will be prepared and submitted to the European Commission.

Pubblicazioni

Peer reviewed articles (5)

FOND planning with explicit fairness assumptions

Autori: Rodriguez, Ivan D.; Bonet, Blai; Sardina, Sebastian; Geffner, Hector
Pubblicato in: Journal of Artificial Intelligence Research, Numero 4, 2022, ISSN 1076-9757
Editore: Morgan Kaufmann Publishers, Inc.
DOI: 10.1613/jair.1.13599

Automated planning instance generation with neuro-symbolic AI

Autori: Carlos Núñez-Molina, Pablo Mesejo, Juan Fernández-Olivares
Pubblicato in: Artificial Intelligence, Numero 352, 2026, Pagina/e 104471, ISSN 0004-3702
Editore: Elsevier BV
DOI: 10.1016/j.artint.2025.104471

First-Order Representation Languages for Goal-Conditioned RL

Autori: Simon Ståhlberg, Hector Geffner
Pubblicato in: Proceedings of the AAAI Conference on Artificial Intelligence, Numero 40, 2026, Pagina/e 36394-36402, ISSN 2374-3468
Editore: Association for the Advancement of Artificial Intelligence (AAAI)
DOI: 10.1609/aaai.v40i43.40960

General Policies, Subgoal Structure, and Planning Width

Autori: Blai Bonet, Hector Geffner
Pubblicato in: Journal of Artificial Intelligence Research, Numero 80, 2024, Pagina/e 475-516, ISSN 1076-9757
Editore: Morgan Kaufmann Publishers, Inc.
DOI: 10.1613/jair.1.15581

Expressing and Exploiting Subgoal Structure in Classical Planning Using Sketches

Autori: Dominik Drexler, Jendrik Seipp, Hector Geffner
Pubblicato in: Journal of Artificial Intelligence Research, Numero 80, 2024, ISSN 1076-9757
Editore: Morgan Kaufmann Publishers, Inc.
DOI: 10.1613/jair.1.15821

Conference proceedings (32)

Learning Generalized Policies Without Supervision Using GNNs

Autori: Simon Stählberg, Blai Bonet, Hector Geffner
Pubblicato in: Proc. Int. Conf. on Principles of Knowledge Representation and Reasoning (KR), 2022
Editore: IJCAI Organization
DOI: 10.48550/arxiv.2205.06002

General Policies, Serializations, and Planning Width

Autori: Bonet, Blai; Geffner, Hector
Pubblicato in: Proceedings AAAI, 2021
Editore: AAAI Press

On Policy Reuse: An Expressive Language for Representing and Executing General Policies that Call Other Policies

Autori: Blai Bonet, Dominik Drexler, Héctor Geffner
Pubblicato in: Proceedings of the International Conference on Automated Planning and Scheduling, Numero 34, 2024, Pagina/e 31-39, ISSN 2334-0843
Editore: AAAI Press
DOI: 10.1609/icaps.v34i1.31458

Expressing and Exploiting the Common Subgoal Structure of Classical Planning Domains Using Sketches.

Autori: Dominik Drexler, Jendrik Seipp, Hector Geffner
Pubblicato in: Proceedings of the 18th International Conference on Principles of Knowledge Representation and Reasoning, 2021
Editore: IJCAI Organization

Learning Generalized Policies for Fully Observable Non-Deterministic Planning Domains

Autori: Hofmann, Till; Geffner, Hector
Pubblicato in: Proc. Int. Joint Conference on Artificial Intelligence (IJCAI), 2025
Editore: IJCAI Organization
DOI: 10.48550/arxiv.2404.02499

Online Action Recognition

Autori: Suárez Hernández, Alejandro; Segovia Aguas, Javier; Torras, Carme; Alenyà Ribas, Guillem
Pubblicato in: Proc. National Conf. on Artificial Intelligence (AAAI), 2021
Editore: AAAI Press
DOI: 10.48550/arxiv.2012.07464

Learning first-order symbolic planning representations that are grounded

Autori: Andres Occhipinti, Blai Bonet, Hector Geffner
Pubblicato in: Proc. ICAPS Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning, 2022
Editore: AAAI Press
DOI: 10.48550/arxiv.2204.11902

Learning General Policies with Policy Gradient Methods

Autori: Simon Ståhlberg, Blai Bonet, Hector Geffner
Pubblicato in: Proceedings of the Twentieth International Conference on Principles of Knowledge Representation and Reasoning, 2023, Pagina/e 647-657, ISBN 978-1-956792-02-7
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2023/63

Learning Generalized Policies for Fully Observable Non-Deterministic Planning Domains

Autori: Till Hofmann, Hector Geffner
Pubblicato in: Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence, 2024, ISBN 978-1-956792-04-1
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/ijcai.2024/744

Learning to Ground Existentially Quantified Goals

Autori: Martin Funkquist, Simon Ståhlberg, Hector Geffner
Pubblicato in: Proc. Int. Conf. on Knowledge Representation and Reasoning, 2024
Editore: IJCAI Org

Learning Lifted STRIPS Models from Action Traces alone: A Simple, General, and Scalable Solution

Autori: Jonas Gösgens, Niklas Jansen, Hector Geffner
Pubblicato in: Proceedings of the International Conference on Automated Planning and Scheduling, Numero 35, 2025, Pagina/e 189-197, ISSN 2334-0843
Editore: Association for the Advancement of Artificial Intelligence (AAAI)
DOI: 10.1609/icaps.v35i1.36117

Learning Sketches for Decomposing Planning Problems into Subproblems of Bounded Width

Autori: Dominik Drexler, Jendrik Seipp, Hector Geffner
Pubblicato in: Proc. Int. Conf. on Planning and Scheduling (ICAPS), 2022
Editore: AAAI Press

Learning Generalized Policies for Fully Observable Non-Deterministic Planning Domains

Autori: Till Hofmann, Hector Geffner
Pubblicato in: Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence, 2025
Editore: IJCAI Organization

Learning General Policies from Small Examples Without Supervision

Autori: Francès, Guillem; Bonet, Blai; Geffner, Hector
Pubblicato in: Proceedings AAAI, 2021
Editore: AAAI Press
DOI: 10.48550/arxiv.2101.00692

Learning to Search and Searching to Learn for Generalization in Planning

Autori: Michael Aichmüller , Yannik Hesse , Hector Geffner
Pubblicato in: Proc. Int. Conf. on Machine Learning (ICML), 2026
Editore: Proc. of Machine Learning Research (PMLR)

Learning More Expressive General Policies for Classical Planning Domains

Autori: Simon Ståhlberg, Blai Bonet, Hector Geffner
Pubblicato in: Proceedings of the AAAI Conference on Artificial Intelligence, Numero 39, 2025, Pagina/e 26697-26706
Editore: Association for the Advancement of Artificial Intelligence (AAAI)
DOI: 10.1609/aaai.v39i25.34872

Symmetries and Expressive Requirements for Learning General Policies

Autori: Dominik Drexler, Simon Ståhlberg, Blai Bonet, Hector Geffner
Pubblicato in: Proc. Int. Conf. on Knowledge Representation and Reasoning, 2024
Editore: IJCAI Org.

Learning Hierarchical Policies by Iteratively Reducing the Width of Sketch Rules

Autori: Dominik Drexler, Jendrik Seipp, Hector Geffner
Pubblicato in: Proceedings of the Twentieth International Conference on Principles of Knowledge Representation and Reasoning, 2023, Pagina/e 208-218
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2023/21

Learning First-Order Representations for Planning from Black Box States: New Results

Autori: Ivan D. Rodriguez; Blai Bonet; Javier Romero; Hector Geffner
Pubblicato in: Proceedings of the 18th International Conference on Principles of Knowledge Representation and Reasoning, 2021
Editore: IJCAI Organization
DOI: 10.24963/kr.2021/51

Generalized Planning as Heuristic Search

Autori: Segovia-Aguas, Javier; Jiménez, Sergio; Jonsson, Anders
Pubblicato in: Proc. Int. Conf. on Planning and Scheduling (ICAPS), 2021
Editore: AAAI Press
DOI: 10.48550/arxiv.2103.14434

Learning Lifted Action Models from Traces of Incomplete Actions and States

Autori: Niklas Jansen, Jonas Gösgens, Hector Geffner
Pubblicato in: Proceedings of the TwentySecond International Conference on Principles of Knowledge Representation and Reasoning, 2025, Pagina/e 832-842
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2025/80

Symmetries and Expressive Requirements for Learning General Policies

Autori: Dominik Drexler, Simon Ståhlberg, Blai Bonet, Hector Geffner
Pubblicato in: Proceedings of the TwentyFirst International Conference on Principles of Knowledge Representation and Reasoning, 2024, Pagina/e 845-855
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2024/79

Flexible FOND Planning with Explicit Fairness Assumptions

Autori: Rodriguez, Ivan D.; Bonet, Blai; Sardina, Sebastian; Geffner, Hector
Pubblicato in: Proceedings of the Thirty-First International Conference on Automated Planning and Scheduling, 2021
Editore: AAAI Press

Learning General Policies from Examples

Autori: Blai Bonet, Hector Geffner
Pubblicato in: Proceedings of the TwentySecond International Conference on Principles of Knowledge Representation and Reasoning, 2025, Pagina/e 740-750
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2025/71

Target Languages (vs. Inductive Biases) for Learning to Act and Plan

Autori: Geffner, Hector
Pubblicato in: Proceedings AAAI, 2022
Editore: AAAI Press

Learning General Optimal Policies with Graph Neural Networks: Expressive Power, Transparency, and Limits

Autori: Ståhlberg, Simon; Bonet, Blai; Geffner, Hector
Pubblicato in: Proceedings of the Thirty-Second International Conference on Automated Planning and Scheduling, 2022
Editore: AAAI Press
DOI: 10.48550/arxiv.2109.10129

From Next Token Prediction to (STRIPS) World Models

Autori: Carlos Núñez-Molina, Vicenç Gómez, Hector Geffner
Pubblicato in: Proc. KR, 2026
Editore: IJCAI Organization

Learning Lifted Action Models from Traces with Minimal Information About Actions and States

Autori: Jonas Gösgens, Niklas Jansen, Hector Geffner
Pubblicato in: Proc. KR., 2026
Editore: IJCAI Organization

Sketch Decompositions for Classical Planning via Deep Reinforcement Learning

Autori: Michael Aichmüller, Hector Geffner
Pubblicato in: Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, 2025, Pagina/e 8438-8446
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/ijcai.2025/938

LTLf Synthesis on First-Order Agent Programs in Nondeterministic Environments

Autori: Till Hofmann, Jens Claßen
Pubblicato in: Proceedings of the AAAI Conference on Artificial Intelligence, Numero 39, 2025, Pagina/e 14976-14986
Editore: Association for the Advancement of Artificial Intelligence (AAAI)
DOI: 10.1609/aaai.v39i14.33642

Combined Task and Motion Planning via Sketch Decompositions

Autori: Magí Dalmau Moreno, Néstor García, Vicenç Gómez, Héctor Geffner
Pubblicato in: Proceedings of the International Conference on Automated Planning and Scheduling, Numero 34, 2024, Pagina/e 123-132, ISSN 2334-0843
Editore: AAAI Press
DOI: 10.1609/icaps.v34i1.31468

Learning to Ground Existentially Quantified Goals

Autori: Martin Funkquist, Simon Ståhlberg, Hector Geffner
Pubblicato in: Proceedings of the TwentyFirst International Conference on Principles of Knowledge Representation and Reasoning, 2024, Pagina/e 856-866
Editore: International Joint Conferences on Artificial Intelligence Organization
DOI: 10.24963/kr.2024/80

Monographic books (1)

Probabilistic and Causal Inference: The Works of Judea Pearl

Autori: Hector Geffner, Rina Dechter, Joseph Y. Halpern (Editors)
Pubblicato in: 2022
Editore: ACM Books
DOI: 10.1145/3501714

Other (2)

Differentiable Learning of Lifted Action Schemas for Classical Planning

Autori: Jonas Reiter, Jakob Elias Gebler, Hector Geffner
Pubblicato in: Submitted, 2026
Editore: Arxiv

Efficient Lookahead Encoding and Abstracted Width for Learning General Policies in Classical Planning

Autori: Michael Aichmüller, Simon Ståhlberg, Martin Funkquist, Hector Geffner
Pubblicato in: Submitted, 2026
Editore: Arxiv