Revisiting Explainable AI through Model-Independent Concept Dictionaries
Abstract
Modern applications of AI rely on increasingly complex models. Explainable AI (XAI) has emerged as a set of techniques aimed at improving model transparency. However, existing XAI methods typically assume input features to be inherently interpretable, or they rely on intermediate internal abstractions that are difficult to characterize and highly architecture-specific, hindering consistent use across models. To address these limitations, we propose DictXAI, a method that defines concepts directly in the input domain via a dictionary---a large, potentially overcomplete set of predefined elements, each carrying an interpretable meaning. Technically, DictXAI first computes a sparse code of the input and then attributes the model's prediction to the associated dictionary elements. We demonstrate the actionable nature of DictXAI explanations, showing that they can attribute AI malfunctions (e.g., Clever Hans effects) directly to identifiable artifact patterns in the data, while fostering human-AI alignment on intricate biomedical signals. We further demonstrate our method's ability to operate across a wide variety of dictionaries, including learned image bases, analytically defined waveforms for electrocardiography, and experimentally acquired dictionary elements. Overall, our results show that DictXAI provides more interpretable, actionable, and architecture-agnostic insights than classical XAI or existing concept-based approaches.
Related
- Extended dynamic mode decomposition with Fourier dictionaries: Error bounds and fast implementation
- Predictions as Surrogates: Revisiting Surrogate Outcomes in the Age of AI
- Revisiting the infinite Galois correspondence via group representations
- Revisiting Fischer decompositions by inframonogenic functions
- Revisiting three-dimensional GLSMs and K-theoretic I-functions
BibTeX
@misc{schnake2026revisiting,
title = {{Revisiting Explainable AI through Model-Independent Concept Dictionaries}},
author = {Thomas Schnake and Doreen Schöppenthau and Alexander Meyer and Jacques Corbeil and Klaus-Robert Müller and Grégoire Montavon},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.10301},
eprint = {2610.10301},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.10301},
}
RIS
TY - UNPB
TI - Revisiting Explainable AI through Model-Independent Concept Dictionaries
AU - Thomas Schnake
AU - Doreen Schöppenthau
AU - Alexander Meyer
AU - Jacques Corbeil
AU - Klaus-Robert Müller
AU - Grégoire Montavon
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.10301
UR - https://arxiv.org/abs/2610.10301
AB - Modern applications of AI rely on increasingly complex models. Explainable AI (XAI) has emerged as a set of techniques aimed at improving model transparency. However, existing XAI methods typically assume input features to be inherently interpretable, or they rely on intermediate internal abstractions that are difficult to characterize and highly architecture-specific, hindering consistent use across models. To address these limitations, we propose DictXAI, a method that defines concepts directly in the input domain via a dictionary---a large, potentially overcomplete set of predefined elements, each carrying an interpretable meaning. Technically, DictXAI first computes a sparse code of the input and then attributes the model's prediction to the associated dictionary elements. We demonstrate the actionable nature of DictXAI explanations, showing that they can attribute AI malfunctions (e.g., Clever Hans effects) directly to identifiable artifact patterns in the data, while fostering human-AI alignment on intricate biomedical signals. We further demonstrate our method's ability to operate across a wide variety of dictionaries, including learned image bases, analytically defined waveforms for electrocardiography, and experimentally acquired dictionary elements. Overall, our results show that DictXAI provides more interpretable, actionable, and architecture-agnostic insights than classical XAI or existing concept-based approaches.
ER -