In Defense of Information Leakage in Concept-based Models

• URL: http://arxiv.org/abs/2606.10669v1
• Date: Tue, 09 Jun 2026 10:19:41 GMT
• Title: In Defense of Information Leakage in Concept-based Models
• Abstract summary: Concept-based models (CMs) learn representations that leak concept-irrelevant information.<n>This leakage is undesirable and should be eradicated as it leads to uninterpretable models.<n>We argue that in real-world settings where concept incompleteness is the norm, some leakage is often necessary for constructing accurate and intervenable CMs.
• Score: 5.959957715491888
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Concept-based models (CMs), deep neural networks that ground their predictions on representations aligned with human-understandable concepts (e.g., "round", "stripes", etc.), have been shown to learn representations that leak concept-irrelevant information. As the traditional narrative goes, this leakage is undesirable and should be eradicated as it leads to uninterpretable models. In this paper, we posit that this conventional view of leakage in CMs is not only ill-posed, as the evidence of how leakage makes a model less interpretable is often inconclusive, but also bound to lead to impractical CMs under common real-world constraints. Specifically, we argue that in real-world settings where concept incompleteness is the norm, some leakage is often necessary for constructing accurate and intervenable CMs. To this end, we propose that there is such a thing as benign leakage and show that, by optimizing a reframing of the typical CM training objective, CMs can encourage and exploit this form of leakage without sacrificing accuracy or intervenability.

Related papers

• DriftScope: Measuring The Hidden Effects of Diffusion Model Adaptation [50.584834483403675]
  Adapting pre-trained text-to-image diffusion models, whether to learn new visual concepts or erase unwanted ones, is routinely evaluated on its intended effects alone.<n>We demonstrate that adaptation systematically damages semantically unrelated concepts in ways that aggregate metrics structurally cannot surface.<n>We introduce DriftScope, a prompt-level diagnostic tool that takes any two model checkpoints and returns a ranked list of tokens whose visual concepts have shifted most between them.
  arXiv  Detail & Related papers  (2026-06-30T20:57:51Z)
• Towards Faithful Multimodal Concept Bottleneck Models [0.07329200485567826]
  Concept Bottleneck Models (CBMs) are interpretable models that route predictions through a layer of human-interpretable concepts.<n>We introduce f-CBM, a multimodal CBM framework built on a vision-language backbone that jointly targets concept detection and leakage mitigation.<n>Experiments demonstrate that f-CBM achieves the best trade-off between task accuracy, concept detection, and leakage reduction.
  arXiv  Detail & Related papers  (2026-03-13T16:56:08Z)
• Rethinking Concept Bottleneck Models: From Pitfalls to Solutions [53.84388497227224]
  Concept Bottleneck Models (CBMs) ground predictions in human-understandable concepts.<n>CBM-Suite is a methodological framework to address these challenges.
  arXiv  Detail & Related papers  (2026-03-05T19:37:49Z)
• Is Model Editing Built on Sand? Revealing Its Illusory Success and Fragile Foundation [50.40861036534546]
  Large language models (LLMs) inevitably encode outdated or incorrect knowledge. Updating, deleting, and forgetting such knowledge is important for alignment, safety, and other issues.<n>To address this issue, model editing has emerged as a promising paradigm: by precisely editing a small subset of parameters such that a specific fact is updated while preserving other knowledge.<n>Despite its great success reported in previous papers, we find the apparent reliability of editing rests on a fragile foundation.<n>Our empirical evidence shows that editing is likely to be based on shortcuts rather than full semantics, calling for an urgent reconsideration of the very basis of model editing before further advancements can
  arXiv  Detail & Related papers  (2025-10-01T07:59:23Z)
• There Was Never a Bottleneck in Concept Bottleneck Models [27.888718857850822]
  Concept Bottleneck Models (CBMs) have emerged as a promising approach to mitigate this issue.<n>We argue that CBMs do not impose a true bottleneck: the fact that a component can predict a concept does not guarantee that it encodes only information about that concept.<n>We propose Minimal Concept Bottleneck Models (MCBMs), which incorporate an Information Bottleneck (IB) objective to constrain each representation component to retain only the information relevant to its corresponding concept.
  arXiv  Detail & Related papers  (2025-06-05T10:50:42Z)
• Seeing Through Deception: Uncovering Misleading Creator Intent in Multimodal News with Vision-Language Models [65.23999399834638]
  We introduce DeceptionDecoded, a benchmark of 12,000 image-caption pairs grounded in trustworthy reference articles.<n>The dataset captures both misleading and non-misleading cases, spanning manipulations across visual and textual modalities.<n>It supports three intent-centric tasks: misleading intent detection, misleading source attribution, and creator desire inference.
  arXiv  Detail & Related papers  (2025-05-21T13:14:32Z)
• Avoiding Leakage Poisoning: Concept Interventions Under Distribution Shifts [10.806525657355872]
  We investigate how concept-based models (CMs) respond to out-of-distribution (OOD) inputs.<n>We introduce MixCEM, a new CM that learns to dynamically exploit leaked information missing from its concepts only when this information is in-distribution.
  arXiv  Detail & Related papers  (2025-04-24T20:24:31Z)
• Leakage and Interpretability in Concept-Based Models [0.24466725954625887]
  Concept Bottleneck Models aim to improve interpretability by predicting high-level intermediate concepts.<n>They are known to suffer from information leakage, whereby models exploit unintended information encoded within the learned concepts.<n>We introduce an information-theoretic framework to rigorously characterise and quantify leakage.
  arXiv  Detail & Related papers  (2025-04-18T22:21:06Z)
• How to Continually Adapt Text-to-Image Diffusion Models for Flexible Customization? [91.49559116493414]
  We propose a novel Concept-Incremental text-to-image Diffusion Model (CIDM) It can resolve catastrophic forgetting and concept neglect to learn new customization tasks in a concept-incremental manner. Experiments validate that our CIDM surpasses existing custom diffusion models.
  arXiv  Detail & Related papers  (2024-10-23T06:47:29Z)
• Tree-Based Leakage Inspection and Control in Concept Bottleneck Models [3.135289953462274]
  Concept Bottleneck Models (CBMs) have gained attention for enhancing interpretability by mapping inputs to intermediate concepts before making final predictions. CBMs often suffer from information leakage, where additional input data, not captured by the concepts, is used to improve task performance. We introduce a novel approach for training both joint and sequential CBMs that allows us to identify and control leakage using decision trees.
  arXiv  Detail & Related papers  (2024-10-08T20:42:19Z)
• Concept Arithmetics for Circumventing Concept Inhibition in Diffusion Models [58.065255696601604]
  We use compositional property of diffusion models, which allows to leverage multiple prompts in a single image generation. We argue that it is essential to consider all possible approaches to image generation with diffusion models that can be employed by an adversary.
  arXiv  Detail & Related papers  (2024-04-21T16:35:16Z)
• Implicit Concept Removal of Diffusion Models [92.55152501707995]
  Text-to-image (T2I) diffusion models often inadvertently generate unwanted concepts such as watermarks and unsafe images. We present the Geom-Erasing, a novel concept removal method based on the geometric-driven control.
  arXiv  Detail & Related papers  (2023-10-09T17:13:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.