# ICON decomposition: Multivariate concept auditing reveals shortcuts that neural networks rely on

arXiv: https://arxiv.org/abs/2608.26083 · HTML: https://mlsanity.com/paper/2608.26083/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 56 references that are in the map.
- First posted on arXiv 26 Aug 2026.

## Builds on (most cited first, 30 of 30)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)](https://mlsanity.com/paper/1711.11279/) (2017)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [Designing and Interpreting Probes with Control Tasks](https://mlsanity.com/paper/1909.03368/) (2019)
- [Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?](https://mlsanity.com/paper/1711.09577/) (2017)
- [BCN20000: Dermoscopic Lesions in the Wild](https://mlsanity.com/paper/1908.02288/) (2019)
- [Probing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance?](https://mlsanity.com/paper/2005.00719/) (2020)
- [Concept-based Explainable Artificial Intelligence: A Survey](https://mlsanity.com/paper/2312.12936/) (2023)
- [Post hoc Explanations may be Ineffective for Detecting Unknown Spurious Correlation](https://mlsanity.com/paper/2212.04629/) (2022)
- [Concept Activation Regions: A Generalized Framework For Concept-Based Explanations](https://mlsanity.com/paper/2209.11222/) (2022)
- [Overlooked factors in concept-based explanations: Dataset choice, concept learnability, and human capability](https://mlsanity.com/paper/2207.09615/) (2022)
- [Regression Concept Vectors for Bidirectional Explanations in Histopathology](https://mlsanity.com/paper/1904.04520/) (2019)
- [Controlling Large Language Models Through Concept Activation Vectors](https://mlsanity.com/paper/2501.05764/) (2025)
- [Navigating Neural Space: Revisiting Concept Activation Vectors to Overcome Directional Divergence](https://mlsanity.com/paper/2202.03482/) (2022)
- [From Hope to Safety: Unlearning Biases of Deep Models via Gradient Penalization in Latent Space](https://mlsanity.com/paper/2308.09437/) (2023)
- [Reveal to Revise: An Explainable AI Life Cycle for Iterative Bias Correction of Deep Models](https://mlsanity.com/paper/2303.12641/) (2023)
- [Explaining Explainability: Recommendations for Effective Use of Concept Activation Vectors](https://mlsanity.com/paper/2404.03713/) (2024)
- [Understanding Inter-Concept Relationships in Concept-Based Models](https://mlsanity.com/paper/2405.18217/) (2024)
- [Evaluating Readability and Faithfulness of Concept-based Explanations](https://mlsanity.com/paper/2404.18533/) (2024)
- [Making Corgis Important for Honeycomb Classification: Adversarial Attacks on Concept-based Explainability Tools](https://mlsanity.com/paper/2110.07120/) (2021)
- [A survey on Concept-based Approaches For Model Improvement](https://mlsanity.com/paper/2403.14566/) (2024)
- [Statistical quantification of confounding bias in predictive modelling](https://mlsanity.com/paper/2111.00814/) (2021)
- [Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research](https://mlsanity.com/paper/2301.08525/) (2023)
- [Deep Nonparametric Conditional Independence Tests for Images](https://mlsanity.com/paper/2411.06140/) (2024)
- [DeepRepViz: Identifying Confounders in Deep Learning Model Predictions](https://mlsanity.com/paper/2309.15551/) (2023)
- [Post-Hoc Concept Disentanglement: From Correlated to Isolated Concept Representations](https://mlsanity.com/paper/2503.05522/) (2025)

## Abstract (arXiv)

Deep neural networks often exploit spurious associations, a failure known as shortcut learning. Before deployment, models should therefore be audited for reliance on concepts such as acquisition artifacts and patient demographics. Current methods, including linear probes and concept activation vectors, measure concept reliance by asking whether each concept is decodable from a layer. We show that this framing is misleading: decodability reflects not only concept reliance but also correlations among concepts and the outcome in the auditing dataset. We reframe shortcut auditing as a decomposition of the variance of intermediary representations with Independent Canonical cONcepts (ICON). ICON quantifies the share of a layer's variance explained by each concept conditional on all other concepts and the outcome, and also reports the variance left unexplained. Because ICON explanations are comparable across layers, their trajectories distinguish concepts that drive predictions from those that are merely encoded in the representations. In simulated data with ground truth, ICON recovers concept importance more accurately than seven baselines. In skin-cancer and neuroimaging applications, it separates learned shortcuts from merely correlated concepts, which we confirm with independent validation checks.