Subgoal Search For Complex Reasoning Tasks
| Title: | Subgoal Search For Complex Reasoning Tasks |
|---|---|
| Authors: | Czechowski, Konrad, Odrzygóźdź, Tomasz, Zbysiński, Marek, Zawalski, Michał, Olejnik, Krzysztof, Wu, Yuhuai, Kuciński, Łukasz, Miłoś, Piotr |
| Publication Year: | 2021 |
| Collection: | Computer Science |
| Subject Terms: | Computer Science - Artificial Intelligence, Computer Science - Machine Learning |
| More Details: | Humans excel in solving complex reasoning tasks through a mental process of moving from one idea to a related one. Inspired by this, we propose Subgoal Search (kSubS) method. Its key component is a learned subgoal generator that produces a diversity of subgoals that are both achievable and closer to the solution. Using subgoals reduces the search space and induces a high-level search graph suitable for efficient planning. In this paper, we implement kSubS using a transformer-based subgoal module coupled with the classical best-first search framework. We show that a simple approach of generating $k$-th step ahead subgoals is surprisingly efficient on three challenging domains: two popular puzzle games, Sokoban and the Rubik's Cube, and an inequality proving benchmark INT. kSubS achieves strong results including state-of-the-art on INT within a modest computational budget. Comment: NeurIPS 2021 |
| Document Type: | Working Paper |
| Access URL: | http://arxiv.org/abs/2108.11204 |
| Accession Number: | edsarx.2108.11204 |
| Database: | arXiv |
| FullText | Text: Availability: 0 CustomLinks: – Url: http://arxiv.org/abs/2108.11204 Name: EDS - Arxiv Category: fullText Text: View this record from Arxiv MouseOverText: View this record from Arxiv – Url: https://resolver.ebsco.com/c/xy5jbn/result?sid=EBSCO:edsarx&genre=article&issn=&ISBN=&volume=&issue=&date=20210825&spage=&pages=&title=Subgoal Search For Complex Reasoning Tasks&atitle=Subgoal%20Search%20For%20Complex%20Reasoning%20Tasks&aulast=Czechowski%2C%20Konrad&id=DOI: Name: Full Text Finder (for New FTF UI) (s8985755) Category: fullText Text: Find It @ SCU Libraries MouseOverText: Find It @ SCU Libraries |
|---|---|
| Header | DbId: edsarx DbLabel: arXiv An: edsarx.2108.11204 RelevancyScore: 1014 AccessLevel: 3 PubType: Report PubTypeId: report PreciseRelevancyScore: 1013.76702880859 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Subgoal Search For Complex Reasoning Tasks – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Czechowski%2C+Konrad%22">Czechowski, Konrad</searchLink><br /><searchLink fieldCode="AR" term="%22Odrzygóźdź%2C+Tomasz%22">Odrzygóźdź, Tomasz</searchLink><br /><searchLink fieldCode="AR" term="%22Zbysiński%2C+Marek%22">Zbysiński, Marek</searchLink><br /><searchLink fieldCode="AR" term="%22Zawalski%2C+Michał%22">Zawalski, Michał</searchLink><br /><searchLink fieldCode="AR" term="%22Olejnik%2C+Krzysztof%22">Olejnik, Krzysztof</searchLink><br /><searchLink fieldCode="AR" term="%22Wu%2C+Yuhuai%22">Wu, Yuhuai</searchLink><br /><searchLink fieldCode="AR" term="%22Kuciński%2C+Łukasz%22">Kuciński, Łukasz</searchLink><br /><searchLink fieldCode="AR" term="%22Miłoś%2C+Piotr%22">Miłoś, Piotr</searchLink> – Name: DatePubCY Label: Publication Year Group: Date Data: 2021 – Name: Subset Label: Collection Group: HoldingsInfo Data: Computer Science – Name: Subject Label: Subject Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Computer+Science+-+Artificial+Intelligence%22">Computer Science - Artificial Intelligence</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+Science+-+Machine+Learning%22">Computer Science - Machine Learning</searchLink> – Name: Abstract Label: Description Group: Ab Data: Humans excel in solving complex reasoning tasks through a mental process of moving from one idea to a related one. Inspired by this, we propose Subgoal Search (kSubS) method. Its key component is a learned subgoal generator that produces a diversity of subgoals that are both achievable and closer to the solution. Using subgoals reduces the search space and induces a high-level search graph suitable for efficient planning. In this paper, we implement kSubS using a transformer-based subgoal module coupled with the classical best-first search framework. We show that a simple approach of generating $k$-th step ahead subgoals is surprisingly efficient on three challenging domains: two popular puzzle games, Sokoban and the Rubik's Cube, and an inequality proving benchmark INT. kSubS achieves strong results including state-of-the-art on INT within a modest computational budget.<br />Comment: NeurIPS 2021 – Name: TypeDocument Label: Document Type Group: TypDoc Data: Working Paper – Name: URL Label: Access URL Group: URL Data: <link linkTarget="URL" linkTerm="http://arxiv.org/abs/2108.11204" linkWindow="_blank">http://arxiv.org/abs/2108.11204</link> – Name: AN Label: Accession Number Group: ID Data: edsarx.2108.11204 |
| PLink | https://login.libproxy.scu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsarx&AN=edsarx.2108.11204 |
| RecordInfo | BibRecord: BibEntity: Subjects: – SubjectFull: Computer Science - Artificial Intelligence Type: general – SubjectFull: Computer Science - Machine Learning Type: general Titles: – TitleFull: Subgoal Search For Complex Reasoning Tasks Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Czechowski, Konrad – PersonEntity: Name: NameFull: Odrzygóźdź, Tomasz – PersonEntity: Name: NameFull: Zbysiński, Marek – PersonEntity: Name: NameFull: Zawalski, Michał – PersonEntity: Name: NameFull: Olejnik, Krzysztof – PersonEntity: Name: NameFull: Wu, Yuhuai – PersonEntity: Name: NameFull: Kuciński, Łukasz – PersonEntity: Name: NameFull: Miłoś, Piotr IsPartOfRelationships: – BibEntity: Dates: – D: 25 M: 08 Type: published Y: 2021 |
| ResultId | 1 |