Alternate social theory discovery using genetic programming

Gunaratne, Chathika; Garibay, Iván

doi:10.1145/3071178.3071332

Cited by 12 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A flow chart depicting the overall process of model discovery is shown in Figure 2. The process both extends and formalises the recent approach described by Gunaratne & Garibay [9]. The different steps in the process are discussed here, making reference to the case study model introduced previously.…”

Section: Model Discovery Methods 31 Overviewmentioning

confidence: 87%

“…More recently, Zhong and colleagues used gene expression programming to optimize the structure of a reward function used by agents to evaluate behavioral choices, such that the ABM could better reproduce empirically observed crowd behaviors [28]. Later, Gunaratne & Garibay used genetic programming to evolve agents' farm selection rules to identify new model structures for a NetLogo implementation of the seminal Artificial Anasazi ABM in order to reproduce the archeological population demography of Long House Valley, Arizona [9]. In all three cases, the focus for structural calibration was on aspects of agent decision-making, rather than any other parts of the model.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Toward inverse generative social science using multi-objective genetic programming

Probst

Epstein

et al. 2019

Proceedings of the Genetic and Evolutionary Computation Conference

View full text Add to dashboard Cite

Generative mechanism-based models of social systems, such as those represented by agent-based simulations, require that intraagent equations (or rules) be specified. However there are often many different choices available for specifying these equations, which can still be interpreted as falling within a particular class of mechanisms. Whilst it is important for a generative model to reproduce historically observed dynamics, it is also important for the model to be theoretically enlightening. Genetic programs (our own included) often produce concatenations that are highly predictive but are complex and hard to interpret theoretically. Here, we develop a new method -based on multi-objective genetic programming -for automating the exploration of both objectives simultaneously. We demonstrate the method by evolving the equations for an existing agent-based simulation of alcohol use behaviors based on social norms theory, the initial model structure for which was developed by a team of human modelers. We discover a trade-off between empirical fit and theoretical interpretability that offers insight into the social norms processes that influence the change and stasis in alcohol use behaviors over time.

show abstract

Section: Model Discovery Methods 31 Overviewmentioning

confidence: 87%

Section: Related Workmentioning

confidence: 99%

Toward inverse generative social science using multi-objective genetic programming

Probst

Epstein

et al. 2019

Proceedings of the Genetic and Evolutionary Computation Conference

View full text Add to dashboard Cite

show abstract

“…Several other studies apply similar structural calibration techniques to various domains, such as opinion dynamics (Husselmann 2015), archaeological simulations (Gunaratne & Garibay 2017), and human crowding models (Junges & Klügl 2011;Zhong et al 2017). In Zhong et al (2017), the authors employ a "dual-layer" architecture consisting of a lower-level social force model for collision avoidance and a higher-level navigation model.…”

Section: 3mentioning

confidence: 99%

Learning Interpretable Logic for Agent-Based Models from Domain Independent Primitives

Greig

Major

Pacholska³

et al. 2023

JASSS

View full text Add to dashboard Cite

Genetic programming (GP) is a powerful method applicable to Inverse Generative Social Science (IGSS) for learning non-trivial agent logic in agent-based models (ABMs). While previous attempts at using evolutionary algorithms for learning ABM structures have focused on recombining domain-specific primitives, this paper extends prior work by developing techniques to evolve interpretable agent logic from scratch using a highly flexible domain-specific language (DSL) comprised of domain-independent primitives, such as basic mathematical operators. We demonstrate the flexibility of our method by learning symbolic models in two distinct domains: flocking and opinion dynamics, targeting data generated by reference models. Our results show that the evolved solutions closely resemble the reference models in behavior, generalize exceptionally well, and exhibit robustness to noise. Additionally, we provide an in-depth analysis of the generated code and intermediate behaviors, revealing the training process's progression. We explore techniques for further enhancing the interpretability of the resulting code and include a population-level analysis of the diversity for both models. This research demonstrates the potential of GP in IGSS for learning interpretable agent logic in ABMs across various domains.

show abstract

“…By contrast, agent-based rules have also been developed for regional land use models, often ad hoc using expert judgment (Thober et al, 2017) informed by empirical studies (Robinson et al, 2007). There remains an opportunity to automate this process via model generation techniques, as has been explored elsewhere in the social sciences (e.g., Gunaratne & Garibay, 2017;Vu et al, 2019). While land use change presents a challenging test case, the methods proposed here also generalize to other aspects of human behavior in water resources systems, contingent on the availability of scale-appropriate datasets.…”

Section: Problem Definitionmentioning

confidence: 99%

“…tem processes (Gunaratne & Garibay, 2017;Schlüter et al, 2019;Vu et al, 2019). By extricating the processes driving emergent and interdependent behaviors in coupled systems, data-driven models can be used beyond the integration of observations to advance theory.…”

mentioning

confidence: 99%

Toward Data‐Driven Generation and Evaluation of Model Structure for Integrated Representations of Human Behavior in Water Resources Systems

Ekblad

Herman

2021

Water Resources Research

View full text Add to dashboard Cite

Simulations of human behavior in water resources systems are challenged by uncertainty in model structure and parameters. The increasing availability of observations describing these systems provides the opportunity to infer a set of plausible model structures using data‐driven approaches. This study develops a three‐phase approach to the inference of model structures and parameterizations from data: problem definition, model generation, and model evaluation, illustrated on a case study of land use decisions in the Tulare Basin, California. We encode the generalized decision problem as an arbitrary mapping from a high‐dimensional data space to the action of interest and use multiobjective genetic programming to search over a family of functions that perform this mapping for both regression and classification tasks. To facilitate the discovery of models that are both realistic and interpretable, the algorithm selects model structures based on multiobjective optimization of (1) their performance on a training set and (2) complexity, measured by the number of variables, constants, and operations composing the model. After training, optimal model structures are further evaluated according to their ability to generalize to held‐out test data and clustered based on their performance, complexity, and generalization properties. Finally, we diagnose the causes of good and bad generalization by performing sensitivity analysis across model inputs and within model clusters. This study serves as a template to inform and automate the problem‐dependent task of constructing robust data‐driven model structures to describe human behavior in water resources systems.

show abstract

Alternate social theory discovery using genetic programming

Cited by 12 publications

References 20 publications

Toward inverse generative social science using multi-objective genetic programming

Toward inverse generative social science using multi-objective genetic programming

Learning Interpretable Logic for Agent-Based Models from Domain Independent Primitives

Toward Data‐Driven Generation and Evaluation of Model Structure for Integrated Representations of Human Behavior in Water Resources Systems

Contact Info

Product

Resources

About