Neurosymbolic Programming for Science
Neurosymbolic Programming (NP) techniques have the potential to accelerate scientific discovery across fields. These models combine neural and symbolic components to learn complex patterns and representations from data, using high-level concepts or known constraints. As a result, NP techniques can i...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | en_US |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/1721.1/145783 |
| _version_ | 1826205318900613120 |
|---|---|
| author | Sun, Jennifer J Tjandrasuwita, Megan Sehgal, Atharva Solar-Lezama, Armando Chaudhuri, Swarat Yue, Yisong Costilla Reyes, Omar |
| author_facet | Sun, Jennifer J Tjandrasuwita, Megan Sehgal, Atharva Solar-Lezama, Armando Chaudhuri, Swarat Yue, Yisong Costilla Reyes, Omar |
| author_sort | Sun, Jennifer J |
| collection | MIT |
| description | Neurosymbolic Programming (NP) techniques have the potential to accelerate scientific discovery across fields. These models combine neural and symbolic components to learn complex patterns and representations from data, using high-level concepts or known constraints. As a result, NP techniques can interface with symbolic domain knowledge from scientists, such as prior knowledge and experimental context, to produce interpretable outputs. Here, we identify opportunities and challenges between current NP models and scientific workflows, with real-world examples from behavior analysis in science. We define concrete next steps to move the NP for science field forward, to enable its use broadly for workflows across the natural and social sciences. |
| first_indexed | 2024-09-23T13:10:48Z |
| format | Article |
| id | mit-1721.1/145783 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T13:10:48Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | mit-1721.1/1457832022-10-12T03:21:48Z Neurosymbolic Programming for Science Sun, Jennifer J Tjandrasuwita, Megan Sehgal, Atharva Solar-Lezama, Armando Chaudhuri, Swarat Yue, Yisong Costilla Reyes, Omar programming languages deep learning science domain knowledge Neurosymbolic Programming (NP) techniques have the potential to accelerate scientific discovery across fields. These models combine neural and symbolic components to learn complex patterns and representations from data, using high-level concepts or known constraints. As a result, NP techniques can interface with symbolic domain knowledge from scientists, such as prior knowledge and experimental context, to produce interpretable outputs. Here, we identify opportunities and challenges between current NP models and scientific workflows, with real-world examples from behavior analysis in science. We define concrete next steps to move the NP for science field forward, to enable its use broadly for workflows across the natural and social sciences. This project was supported by the the National Science Foundation under Grant No. 1918839 "Understanding the World Through Code" http://www.neurosymbolic.org 2022-10-12T01:58:33Z 2022-10-12T01:58:33Z 2022-10-12 Article https://hdl.handle.net/1721.1/145783 en_US application/pdf |
| spellingShingle | programming languages deep learning science domain knowledge Sun, Jennifer J Tjandrasuwita, Megan Sehgal, Atharva Solar-Lezama, Armando Chaudhuri, Swarat Yue, Yisong Costilla Reyes, Omar Neurosymbolic Programming for Science |
| title | Neurosymbolic Programming for Science |
| title_full | Neurosymbolic Programming for Science |
| title_fullStr | Neurosymbolic Programming for Science |
| title_full_unstemmed | Neurosymbolic Programming for Science |
| title_short | Neurosymbolic Programming for Science |
| title_sort | neurosymbolic programming for science |
| topic | programming languages deep learning science domain knowledge |
| url | https://hdl.handle.net/1721.1/145783 |
| work_keys_str_mv | AT sunjenniferj neurosymbolicprogrammingforscience AT tjandrasuwitamegan neurosymbolicprogrammingforscience AT sehgalatharva neurosymbolicprogrammingforscience AT solarlezamaarmando neurosymbolicprogrammingforscience AT chaudhuriswarat neurosymbolicprogrammingforscience AT yueyisong neurosymbolicprogrammingforscience AT costillareyesomar neurosymbolicprogrammingforscience |