Julien Colin

Thomas Fel, Agustin Picard, Louis Bethune et al. · harvard

Attribution methods, which employ heatmaps to identify the most influential regions of an image that impact model decisions, have gained widespread popularity as a type of explainability method. However, recent research has exposed the limited practical value of these methods, attributed in part to their narrow focus on the most prominent regions of an image -- revealing "where" the model looks, but failing to elucidate "what" the model sees in those areas. In this work, we try to fill in this gap with CRAFT -- a novel approach to identify both "what" and "where" by generating concept-based explanations. We introduce 3 new ingredients to the automatic concept extraction literature: (i) a recursive strategy to detect and decompose concepts across layers, (ii) a novel method for a more faithful estimation of concept importance using Sobol indices, and (iii) the use of implicit differentiation to unlock Concept Attribution Maps. We conduct both human and computer vision experiments to demonstrate the benefits of the proposed approach. We show that the proposed concept importance estimation technique is more faithful to the model than previous methods. When evaluating the usefulness of the method for human experimenters on a human-centered utility benchmark, we find that our approach significantly improves on two of the three test scenarios. Our code is freely available at github.com/deel-ai/Craft.

Thomas Fel, Lucas Hervier, David Vigouroux et al. · harvard

Today's most advanced machine-learning models are hardly scrutable. The key challenge for explainability methods is to help assisting researchers in opening up these black boxes, by revealing the strategy that led to a given decision, by characterizing their internal states or by studying the underlying data representation. To address this challenge, we have developed Xplique: a software library for explainability which includes representative explainability methods as well as associated evaluation metrics. It interfaces with one of the most popular learning libraries: Tensorflow as well as other libraries including PyTorch, scikit-learn and Theano. The code is licensed under the MIT license and is freely available at github.com/deel-ai/xplique.

Thomas Fel, Thibaut Boissin, Victor Boutin et al. · harvard

Feature visualization has gained substantial popularity, particularly after the influential work by Olah et al. in 2017, which established it as a crucial tool for explainability. However, its widespread adoption has been limited due to a reliance on tricks to generate interpretable images, and corresponding challenges in scaling it to deeper neural networks. Here, we describe MACO, a simple approach to address these shortcomings. The main idea is to generate images by optimizing the phase spectrum while keeping the magnitude constant to ensure that generated explanations lie in the space of natural images. Our approach yields significantly better results (both qualitatively and quantitatively) and unlocks efficient and interpretable feature visualizations for large state-of-the-art neural networks. We also show that our approach exhibits an attribution mechanism allowing us to augment feature visualizations with spatial importance. We validate our method on a novel benchmark for comparing feature visualization methods, and release its visualizations for all classes of the ImageNet dataset on https://serre-lab.github.io/Lens/. Overall, our approach unlocks, for the first time, feature visualizations for large, state-of-the-art deep neural networks without resorting to any parametric prior image model.

Victor Boutin, Thomas Fel, Lakshya Singhal et al. · harvard

An important milestone for AI is the development of algorithms that can produce drawings that are indistinguishable from those of humans. Here, we adapt the 'diversity vs. recognizability' scoring framework from Boutin et al, 2022 and find that one-shot diffusion models have indeed started to close the gap between humans and machines. However, using a finer-grained measure of the originality of individual samples, we show that strengthening the guidance of diffusion models helps improve the humanness of their drawings, but they still fall short of approximating the originality and recognizability of human drawings. Comparing human category diagnostic features, collected through an online psychophysics experiment, against those derived from diffusion models reveals that humans rely on fewer and more localized features. Overall, our study suggests that diffusion models have significantly helped improve the quality of machine-generated drawings; however, a gap between humans and machines remains -- in part explainable by discrepancies in visual strategies.

Aimen Zerroug, Mohit Vaishnav, Julien Colin et al.

A fundamental component of human vision is our ability to parse complex visual scenes and judge the relations between their constituent objects. AI benchmarks for visual reasoning have driven rapid progress in recent years with state-of-the-art systems now reaching human accuracy on some of these benchmarks. Yet, a major gap remains in terms of the sample efficiency with which humans and AI systems learn new visual reasoning tasks. Humans' remarkable efficiency at learning has been at least partially attributed to their ability to harness compositionality -- such that they can efficiently take advantage of previously gained knowledge when learning new tasks. Here, we introduce a novel visual reasoning benchmark, Compositional Visual Relations (CVR), to drive progress towards the development of more data-efficient learning algorithms. We take inspiration from fluidic intelligence and non-verbal reasoning tests and describe a novel method for creating compositions of abstract rules and associated image datasets at scale. Our proposed benchmark includes measures of sample efficiency, generalization and transfer across task rules, as well as the ability to leverage compositionality. We systematically evaluate modern neural architectures and find that, surprisingly, convolutional architectures surpass transformer-based architectures across all performance measures in most data regimes. However, all computational models are a lot less data efficient compared to humans even after learning informative visual representations using self-supervision. Overall, we hope that our challenge will spur interest in the development of neural architectures that can learn to harness compositionality toward more efficient learning.

Julien Colin, Lore Goetschalckx, Nuria Oliver et al.

How interpretable are the features of leading vision models? The question is increasingly pressing as these models move from research benchmarks into high-stakes deployments, yet existing methods cannot answer it reliably. We close this gap with a framework for measuring and comparing the human interpretability of vision models, built around two complementary psychophysics protocols: (1) localizability -- can an observer predict where a feature fires on a novel image? -- and (2) nameability -- can an observer accurately describe what the feature represents? Features are recovered via sparse autoencoders, and a chance-anchored scoring function places every model on a common scale. Applying the framework to six vision transformers -- two supervised ViTs and four foundation models (DINOv2, DINOv3, CLIP, SigLIP) -- we collected more than $15{,}000$ behavioral responses, analyzing the $13{,}400$ responses from the $377$ participants who passed our pre-specified quality checks. Foundation models are consistently *less* interpretable than their supervised counterparts, and the gap is not a capability tradeoff: interpretability does not correlate with downstream task performance on any benchmark we examine. What does correlate is the locality of a feature's activations and coarse-grained semantic alignment with humans -- models with focal activations and representations that reflect the world's broad categorical structure produce more interpretable features, whereas fine-grained perceptual alignment does not. The two protocols yield strongly correlated rankings and share the same predictors, establishing interpretability as an independent, measurable dimension of representation quality -- and, surprisingly, one on which every foundation model we tested falls below the supervised baselines that came before. Capability alone cannot close that gap; locality and coarse-grained alignment can.

Julien Colin, Lore Goetschalckx, Thomas Fel et al. · harvard

Much of the research on the interpretability of deep neural networks has focused on studying the visual features that maximally activate individual neurons. However, recent work has cast doubts on the usefulness of such local representations for understanding the behavior of deep neural networks because individual neurons tend to respond to multiple unrelated visual patterns, a phenomenon referred to as "superposition". A promising alternative to disentangle these complex patterns is learning sparsely distributed vector representations from entire network layers, as the resulting basis vectors seemingly encode single identifiable visual patterns consistently. Thus, one would expect the resulting code to align better with human perceivable visual patterns, but supporting evidence remains, at best, anecdotal. To fill this gap, we conducted three large-scale psychophysics experiments collected from a pool of 560 participants. Our findings provide (i) strong evidence that features obtained from sparse distributed representations are easier to interpret by human observers and (ii) that this effect is more pronounced in the deepest layers of a neural network. Complementary analyses also reveal that (iii) features derived from sparse distributed representations contribute more to the model's decision. Overall, our results highlight that distributed representations constitute a superior basis for interpretability, underscoring a need for the field to move beyond the interpretation of local neural codes in favor of sparsely distributed ones.

Julien Colin, Thomas Fel, Remi Cadene et al.

A multitude of explainability methods and associated fidelity performance metrics have been proposed to help better understand how modern AI systems make decisions. However, much of the current work has remained theoretical -- without much consideration for the human end-user. In particular, it is not yet known (1) how useful current explainability methods are in practice for more real-world scenarios and (2) how well associated performance metrics accurately predict how much knowledge individual explanations contribute to a human end-user trying to understand the inner-workings of the system. To fill this gap, we conducted psychophysics experiments at scale to evaluate the ability of human participants to leverage representative attribution methods for understanding the behavior of different image classifiers representing three real-world scenarios: identifying bias in an AI system, characterizing the visual strategy it uses for tasks that are too difficult for an untrained non-expert human observer as well as understanding its failure cases. Our results demonstrate that the degree to which individual attribution methods help human participants better understand an AI system varied widely across these scenarios. This suggests a critical need for the field to move past quantitative improvements of current attribution methods towards the development of complementary approaches that provide qualitatively different sources of information to human end-users.