Multiplicity of Artificial Intelligence Models and Its Importance in Bias Control

Autores/as

DOI:

https://doi.org/10.26439/ciii2025.8662

Resumen

The so-called Rashomon effect is inspired by Akira Kurosawa’s film Rashomon. In the film, four witnesses present different and contradictory accounts of the same crime. A similar situation is observed when multiple artificial intelligence (AI) and machine learning (ML) models generate solutions that achieve the same level of performance yet differ from one another when applied to the same dataset.

Descargas

Los datos de descarga aún no están disponibles.

Biografía del autor/a

  • Juan R. Jaramillo, Long Island University, New York, USA

    Juan Jaramillo holds a PhD in Industrial Engineering from West Virginia University. He is currently an Associate Professor of Data Analytics at Long Island University. He also serves as President of the INFORMS Analytics Society and as an Advisory Board Member for Dhaxle. He has co-authored “The Innovation Sandbox: Why Organización Corona Partners with Academia to Solve the Data Challenges of Tomorrow” (OR/MS Today, 2026); “A Deep Learning Counting Model Applied to Quality Control” (Journal of Modelling in Management, 2023); “Alcohol, Illicit Drugs, and Suicide Mortality Trends Stratified by Age, Gender, and Race for 2006–2019” (American Journal of Lifestyle Medicine, 2023); and the text book Machine Learning for Business Analytics (Routledge, 2022). His research interests include machine learning, deep learning, bias control in AI, supply chain analytics, and the application of data envelopment analysis to healthcare and economic performance.

Referencias

[1] L. Breiman, “Statistical modeling: The two cultures,” Stat. Sci., vol. 16, no. 3, pp. 199–215, Aug. 2001, doi: 10.1214/ss/1009213726

[2] C. Rudin et al., “Amazing things come from having many good models,” 2025, arXiv:2407.04846.

[3] L. Semenova, C. Rudin, and R. Parr, “On the existence of simpler machine learning models,” in Proc. ACM Conf. Fairness, Accountability, and Transparency (FAccT ’22), Seoul, Republic of Korea, 2022, pp. 1827–1858, doi: 10.1145/3531146.3533232.

[4] Fisher, C. Rudin, and F. Dominici, “All models are wrong, but many are useful: Learning a variable’s importance by studying an entire class of prediction models simultaneously,” J. Mach. Learn. Res., vol. 20, no. 177, pp. 1–81, Jan. 2019. [Online]. Available: https://jmlr.csail.mit.edu/papers/volume20/18-760/18-760.pdf

Descargas

Publicado

2026-06-08

Número

Núm. 004 (2025): IV Congreso Internacional de Ingeniería Industrial

Sección

Ponencias

Licencia

Derechos de autor 2026 Congreso Internacional de Ingeniería Industrial de la Universidad de Lima

Creative Commons License

Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.