Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains. Staud, M. In Proceedings of the 6th ICAPS Workshop on Hierarchical Planning (HPlan 2023), pages 19–27, 2023. ![Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper abstract bibtex In this paper, we present a novel approach that combines Hierarchical Task Planning (HTN) with deep learning techniques to address the challenges of scalability and efficiency in large-scale planning problems. Building upon the Hierarchical World State Planning (HWSP) algorithm, our method utilizes a multi-layered world state representation, which allows for planning at abstract levels without the need to consider lower-level details. We propose a deep learning method for predicting the effects of abstract tasks, which opens the door to enhancements in both planning performance and plan quality. Additionally, we employ the same approach to create a domain-dependent planning heuristic. Our contributions demonstrate the potential of integrating HTN planning with deep learning techniques, paving the way for future research in various application domains such as robotics, logistics, and urban planning. The proposed approach employs standard deep learning techniques, ensuring adaptability as the state of the art advances.
@InProceedings{Staud2023DLandHTNsFor2D,
author = {Michael Staud},
title = {Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains},
booktitle = {Proceedings of the 6th ICAPS Workshop on Hierarchical Planning (HPlan 2023)},
year = {2023},
pages = {19--27},
abstract = {In this paper, we present a novel approach that combines Hierarchical Task Planning (HTN) with deep learning techniques to address the challenges of scalability and efficiency in large-scale planning problems. Building upon the Hierarchical World State Planning (HWSP) algorithm, our method utilizes a multi-layered world state representation, which allows for planning at abstract levels without the need to consider lower-level details. We propose a deep learning method for predicting the effects of abstract tasks, which opens the door to enhancements in both planning performance and plan quality. Additionally, we employ the same approach to create a domain-dependent planning heuristic. Our contributions demonstrate the potential of integrating HTN planning with deep learning techniques, paving the way for future research in various application domains such as robotics, logistics, and urban planning. The proposed approach employs standard deep learning techniques, ensuring adaptability as the state of the art advances.},
url_paper = {https://icaps23.icaps-conference.org/papers/hplan/HPlan2023_paper_7.pdf}
}
Downloads: 0
{"_id":"5L5v263NkyEtjRgZk","bibbaseid":"staud-integratingdeeplearningtechniquesintohierarchicaltaskplanningforeffectandheuristicpredictionsin2ddomains-2023","author_short":["Staud, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["Michael"],"propositions":[],"lastnames":["Staud"],"suffixes":[]}],"title":"Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains","booktitle":"Proceedings of the 6th ICAPS Workshop on Hierarchical Planning (HPlan 2023)","year":"2023","pages":"19–27","abstract":"In this paper, we present a novel approach that combines Hierarchical Task Planning (HTN) with deep learning techniques to address the challenges of scalability and efficiency in large-scale planning problems. Building upon the Hierarchical World State Planning (HWSP) algorithm, our method utilizes a multi-layered world state representation, which allows for planning at abstract levels without the need to consider lower-level details. We propose a deep learning method for predicting the effects of abstract tasks, which opens the door to enhancements in both planning performance and plan quality. Additionally, we employ the same approach to create a domain-dependent planning heuristic. Our contributions demonstrate the potential of integrating HTN planning with deep learning techniques, paving the way for future research in various application domains such as robotics, logistics, and urban planning. The proposed approach employs standard deep learning techniques, ensuring adaptability as the state of the art advances.","url_paper":"https://icaps23.icaps-conference.org/papers/hplan/HPlan2023_paper_7.pdf","bibtex":"@InProceedings{Staud2023DLandHTNsFor2D,\n author = {Michael Staud},\n title = {Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains},\n booktitle = {Proceedings of the 6th ICAPS Workshop on Hierarchical Planning (HPlan 2023)},\n year = {2023},\n pages = {19--27},\n abstract = {In this paper, we present a novel approach that combines Hierarchical Task Planning (HTN) with deep learning techniques to address the challenges of scalability and efficiency in large-scale planning problems. Building upon the Hierarchical World State Planning (HWSP) algorithm, our method utilizes a multi-layered world state representation, which allows for planning at abstract levels without the need to consider lower-level details. We propose a deep learning method for predicting the effects of abstract tasks, which opens the door to enhancements in both planning performance and plan quality. Additionally, we employ the same approach to create a domain-dependent planning heuristic. Our contributions demonstrate the potential of integrating HTN planning with deep learning techniques, paving the way for future research in various application domains such as robotics, logistics, and urban planning. The proposed approach employs standard deep learning techniques, ensuring adaptability as the state of the art advances.},\n url_paper = {https://icaps23.icaps-conference.org/papers/hplan/HPlan2023_paper_7.pdf}\n}\n\n","author_short":["Staud, M."],"key":"Staud2023DLandHTNsFor2D","id":"Staud2023DLandHTNsFor2D","bibbaseid":"staud-integratingdeeplearningtechniquesintohierarchicaltaskplanningforeffectandheuristicpredictionsin2ddomains-2023","role":"author","urls":{" paper":"https://icaps23.icaps-conference.org/papers/hplan/HPlan2023_paper_7.pdf"},"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"hierarchical-task.net/bibtex/hplan-publications.bib","dataSources":["6h6sNir9HQYRtQmou","T9LQLt3D2MDhrCfjh"],"keywords":[],"search_terms":["integrating","deep","learning","techniques","hierarchical","task","planning","effect","heuristic","predictions","domains","staud"],"title":"Integrating Deep Learning Techniques into Hierarchical Task Planning for Effect and Heuristic Predictions in 2D Domains","year":2023,"downloads":1}