AI-assisted discovery of quantitative and formal models in social science
Abstract In social science, formal and quantitative models, ranging from ones that describe economic growth to collective action, are used to formulate mechanistic explanations of the observed phenomena, provide predictions, and uncover new research questions. Here, we demonstrate the use of a machi...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer Nature
2025-01-01
|
| Series: | Humanities & Social Sciences Communications |
| Online Access: | https://doi.org/10.1057/s41599-025-04405-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832571930797932544 |
|---|---|
| author | Julia Balla Sihao Huang Owen Dugan Rumen Dangovski Marin Soljačić |
| author_facet | Julia Balla Sihao Huang Owen Dugan Rumen Dangovski Marin Soljačić |
| author_sort | Julia Balla |
| collection | DOAJ |
| description | Abstract In social science, formal and quantitative models, ranging from ones that describe economic growth to collective action, are used to formulate mechanistic explanations of the observed phenomena, provide predictions, and uncover new research questions. Here, we demonstrate the use of a machine learning system to aid the discovery of symbolic models that capture non-linear and dynamical relationships in social science datasets. By extending neuro-symbolic methods to find compact functions and differential equations in noisy and longitudinal data, we show that our system can be used to discover interpretable models from real-world data in economics and sociology. Augmenting existing workflows with symbolic regression can help uncover novel relationships and explore counterfactual models during the scientific process. We propose that this AI-assisted framework can bridge parametric and non-parametric models commonly employed in social science research by systematically exploring the space of non-linear models and enabling fine-grained control over expressivity and interpretability. |
| format | Article |
| id | doaj-art-64fba0b7008e485fbd66f8af2e6377b1 |
| institution | Kabale University |
| issn | 2662-9992 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Springer Nature |
| record_format | Article |
| series | Humanities & Social Sciences Communications |
| spelling | doaj-art-64fba0b7008e485fbd66f8af2e6377b12025-02-02T12:13:12ZengSpringer NatureHumanities & Social Sciences Communications2662-99922025-01-0112111210.1057/s41599-025-04405-xAI-assisted discovery of quantitative and formal models in social scienceJulia Balla0Sihao Huang1Owen Dugan2Rumen Dangovski3Marin Soljačić4Department of Electrical Engineering and Computer Science, Massachusetts Institute of TechnologyDepartment of Politics and International Relations, University of OxfordNSF AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI)Department of Electrical Engineering and Computer Science, Massachusetts Institute of TechnologyNSF AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI)Abstract In social science, formal and quantitative models, ranging from ones that describe economic growth to collective action, are used to formulate mechanistic explanations of the observed phenomena, provide predictions, and uncover new research questions. Here, we demonstrate the use of a machine learning system to aid the discovery of symbolic models that capture non-linear and dynamical relationships in social science datasets. By extending neuro-symbolic methods to find compact functions and differential equations in noisy and longitudinal data, we show that our system can be used to discover interpretable models from real-world data in economics and sociology. Augmenting existing workflows with symbolic regression can help uncover novel relationships and explore counterfactual models during the scientific process. We propose that this AI-assisted framework can bridge parametric and non-parametric models commonly employed in social science research by systematically exploring the space of non-linear models and enabling fine-grained control over expressivity and interpretability.https://doi.org/10.1057/s41599-025-04405-x |
| spellingShingle | Julia Balla Sihao Huang Owen Dugan Rumen Dangovski Marin Soljačić AI-assisted discovery of quantitative and formal models in social science Humanities & Social Sciences Communications |
| title | AI-assisted discovery of quantitative and formal models in social science |
| title_full | AI-assisted discovery of quantitative and formal models in social science |
| title_fullStr | AI-assisted discovery of quantitative and formal models in social science |
| title_full_unstemmed | AI-assisted discovery of quantitative and formal models in social science |
| title_short | AI-assisted discovery of quantitative and formal models in social science |
| title_sort | ai assisted discovery of quantitative and formal models in social science |
| url | https://doi.org/10.1057/s41599-025-04405-x |
| work_keys_str_mv | AT juliaballa aiassisteddiscoveryofquantitativeandformalmodelsinsocialscience AT sihaohuang aiassisteddiscoveryofquantitativeandformalmodelsinsocialscience AT owendugan aiassisteddiscoveryofquantitativeandformalmodelsinsocialscience AT rumendangovski aiassisteddiscoveryofquantitativeandformalmodelsinsocialscience AT marinsoljacic aiassisteddiscoveryofquantitativeandformalmodelsinsocialscience |