Toward A Causal Framework for Modeling Perception
Jose M. Alvarez, Salvatore Ruggieri
TL;DR
Perception leads to divergent interpretations of the same ML-provided information in decision flows, challenging reliable automated and human-in-the-loop systems. The authors propose a two-layer framework: first a probabilistic perception over $P(\mathbf{X})$, then a causal perception built on structural causal models $\mathcal{M}$ with $X_j := f_j(X_{pa(j)}, U_j)$ and interventional distributions $P(\mathbf{X})^{do(i)}$ to capture what-if reasoning. They distinguish structural and parametrical forms of causal perception, and propose a mechanism by which individuals construct their own SCMs via categorization and signification, leading to observable disagreement in both observational and interventional distributions. The framework also connects to fairness concerns by addressing situated biases and loaded attributes, and suggests directions for empirical validation and ontology-based SCM construction.
Abstract
Perception occurs when individuals interpret the same information differently. It is a known cognitive phenomenon with implications for bias in human decision-making. Perception, however, remains understudied in machine learning (ML). This is problematic as modern decision flows, whether partially or fully automated by ML applications, always involve human experts. For instance, how might we account for cases in which two experts interpret differently the same deferred instance or explanation from a ML model? Addressing this and similar questions requires first a formulation of perception, particularly, in a manner that integrates with ML-enabled decision flows. In this work, we present a first approach to modeling perception causally. We define perception under causal reasoning using structural causal models (SCMs). Our approach formalizes individual experience as additional causal knowledge that comes with and is used by the expert decision-maker in the form of a SCM. We define two kinds of probabilistic causal perception: structural and parametrical. We showcase our framework through a series of examples of modern decision flows. We also emphasize the importance of addressing perception in fair ML, discussing relevant fairness implications and possible applications.