All AI Models are Wrong, but Some are Optimal

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Anand, Akhil S, Sawant, Shambhuraj, Reinhardt, Dirk, Gros, Sebastien
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910779720073216
author Anand, Akhil S
Sawant, Shambhuraj
Reinhardt, Dirk
Gros, Sebastien
author_facet Anand, Akhil S
Sawant, Shambhuraj
Reinhardt, Dirk
Gros, Sebastien
contents AI models that predict the future behavior of a system (a.k.a. predictive AI models) are central to intelligent decision-making. However, decision-making using predictive AI models often results in suboptimal performance. This is primarily because AI models are typically constructed to best fit the data, and hence to predict the most likely future rather than to enable high-performance decision-making. The hope that such prediction enables high-performance decisions is neither guaranteed in theory nor established in practice. In fact, there is increasing empirical evidence that predictive models must be tailored to decision-making objectives for performance. In this paper, we establish formal (necessary and sufficient) conditions that a predictive model (AI-based or not) must satisfy for a decision-making policy established using that model to be optimal. We then discuss their implications for building predictive AI models for sequential decision-making.
format Preprint
id arxiv_https___arxiv_org_abs_2501_06086
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle All AI Models are Wrong, but Some are Optimal
Anand, Akhil S
Sawant, Shambhuraj
Reinhardt, Dirk
Gros, Sebastien
Artificial Intelligence
Machine Learning
AI models that predict the future behavior of a system (a.k.a. predictive AI models) are central to intelligent decision-making. However, decision-making using predictive AI models often results in suboptimal performance. This is primarily because AI models are typically constructed to best fit the data, and hence to predict the most likely future rather than to enable high-performance decision-making. The hope that such prediction enables high-performance decisions is neither guaranteed in theory nor established in practice. In fact, there is increasing empirical evidence that predictive models must be tailored to decision-making objectives for performance. In this paper, we establish formal (necessary and sufficient) conditions that a predictive model (AI-based or not) must satisfy for a decision-making policy established using that model to be optimal. We then discuss their implications for building predictive AI models for sequential decision-making.
title All AI Models are Wrong, but Some are Optimal
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2501.06086