Temporal and Object Quantification Networks
<jats:p>We present Temporal and Object Quantification Networks (TOQ-Nets), a new class of neuro-symbolic networks with a structural bias that enables them to learn to recognize complex relational-temporal events. This is done by including reasoning layers that implement finite-domain quantific...
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
International Joint Conferences on Artificial Intelligence
2022
|
| Online Access: | https://hdl.handle.net/1721.1/143777 |
| _version_ | 1811073572173512704 |
|---|---|
| author | Mao, Jiayuan Luo, Zhezheng Gan, Chuang Tenenbaum, Joshua B Wu, Jiajun Kaelbling, Leslie Pack Ullman, Tomer D |
| author2 | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences |
| author_facet | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Mao, Jiayuan Luo, Zhezheng Gan, Chuang Tenenbaum, Joshua B Wu, Jiajun Kaelbling, Leslie Pack Ullman, Tomer D |
| author_sort | Mao, Jiayuan |
| collection | MIT |
| description | <jats:p>We present Temporal and Object Quantification Networks (TOQ-Nets), a new class of neuro-symbolic networks with a structural bias that enables them to learn to recognize complex relational-temporal events. This is done by including reasoning layers that implement finite-domain quantification over objects and time. The structure allows them to generalize directly to input instances with varying numbers of objects in temporal sequences of varying lengths. We evaluate TOQ-Nets on input domains that require recognizing event-types in terms of complex temporal relational patterns. We demonstrate that TOQ-Nets can generalize from small amounts of data to scenarios containing more objects than were present during training and to temporal warpings of input sequences.</jats:p> |
| first_indexed | 2024-09-23T09:35:06Z |
| format | Article |
| id | mit-1721.1/143777 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T09:35:06Z |
| publishDate | 2022 |
| publisher | International Joint Conferences on Artificial Intelligence |
| record_format | dspace |
| spelling | mit-1721.1/1437772023-03-30T20:45:26Z Temporal and Object Quantification Networks Mao, Jiayuan Luo, Zhezheng Gan, Chuang Tenenbaum, Joshua B Wu, Jiajun Kaelbling, Leslie Pack Ullman, Tomer D Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science MIT-IBM Watson AI Lab <jats:p>We present Temporal and Object Quantification Networks (TOQ-Nets), a new class of neuro-symbolic networks with a structural bias that enables them to learn to recognize complex relational-temporal events. This is done by including reasoning layers that implement finite-domain quantification over objects and time. The structure allows them to generalize directly to input instances with varying numbers of objects in temporal sequences of varying lengths. We evaluate TOQ-Nets on input domains that require recognizing event-types in terms of complex temporal relational patterns. We demonstrate that TOQ-Nets can generalize from small amounts of data to scenarios containing more objects than were present during training and to temporal warpings of input sequences.</jats:p> 2022-07-15T17:22:16Z 2022-07-15T17:22:16Z 2021 2022-07-15T17:13:47Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/143777 Mao, Jiayuan, Luo, Zhezheng, Gan, Chuang, Tenenbaum, Joshua B, Wu, Jiajun et al. 2021. "Temporal and Object Quantification Networks." Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. en 10.24963/IJCAI.2021/386 Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf International Joint Conferences on Artificial Intelligence MIT web domain |
| spellingShingle | Mao, Jiayuan Luo, Zhezheng Gan, Chuang Tenenbaum, Joshua B Wu, Jiajun Kaelbling, Leslie Pack Ullman, Tomer D Temporal and Object Quantification Networks |
| title | Temporal and Object Quantification Networks |
| title_full | Temporal and Object Quantification Networks |
| title_fullStr | Temporal and Object Quantification Networks |
| title_full_unstemmed | Temporal and Object Quantification Networks |
| title_short | Temporal and Object Quantification Networks |
| title_sort | temporal and object quantification networks |
| url | https://hdl.handle.net/1721.1/143777 |
| work_keys_str_mv | AT maojiayuan temporalandobjectquantificationnetworks AT luozhezheng temporalandobjectquantificationnetworks AT ganchuang temporalandobjectquantificationnetworks AT tenenbaumjoshuab temporalandobjectquantificationnetworks AT wujiajun temporalandobjectquantificationnetworks AT kaelblinglesliepack temporalandobjectquantificationnetworks AT ullmantomerd temporalandobjectquantificationnetworks |