Human-interpretable model explainability on high-dimensional data

Damien de Mijolla, Christopher Frye, Markus Kunesch, John Mansir, Ilya Feige

Abstract: The importance of explainability in machine learning continues to grow, as both neural-network architectures and the data they model become increasingly complex. Unique challenges arise when a model's input features become high-dimensional: on one hand, principled model-agnostic approaches to explainability become too computationally expensive; on the other, more efficient explainability algorithms lack natural interpretations for general users. In this work, we introduce a framework for human-interpretable explainability on high-dimensional data, consisting of two modules. First, we apply a semantically-meaningful latent representation, both to reduce the raw dimensionality of the data, and to ensure its human interpretability. These latent features can be learnt, e.g. explicitly as disentangled representations or implicitly through image-to-image translation, or they can be based on any computable quantities the user chooses. Second, we adapt the Shapley paradigm for model-agnostic explainability to operate on these latent features. This leads to interpretable model explanations that are both theoretically controlled and computationally tractable. We benchmark our approach on synthetic data and demonstrate its effectiveness on several image-classification tasks.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: We adapt Shapley explainability to operate on semantic latent features in order to produce human-interpretable model explanations.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2010.07384/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=r_HG5hHVSb