Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks’ Internal Representations

Aditya Taparia, Som Sagar, Ransalu Senanayake
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:59154-59181, 2025.

Abstract

Understanding the inner representation of a neural network helps users improve models. Concept-based methods have become a popular choice for explaining deep neural networks post-hoc because, unlike most other explainable AI techniques, they can be used to test high-level visual "concepts" that are not directly related to feature attributes. For instance, the concept of "stripes" is important to classify an image as a zebra. Concept-based explanation methods, however, require practitioners to guess and manually collect multiple candidate concept image sets, making the process labor-intensive and prone to overlooking important concepts. Addressing this limitation, in this paper, we frame concept image set creation as an image generation problem. However, since naively using a standard generative model does not result in meaningful concepts, we devise a reinforcement learning-based preference optimization (RLPO) algorithm that fine-tunes a vision-language generative model from approximate textual descriptions of concepts. Through a series of experiments, we demonstrate our method’s ability to efficiently and reliably articulate diverse concepts that are otherwise challenging to craft manually.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-taparia25a, title = {Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks’ Internal Representations}, author = {Taparia, Aditya and Sagar, Som and Senanayake, Ransalu}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {59154--59181}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/taparia25a/taparia25a.pdf}, url = {https://proceedings.mlr.press/v267/taparia25a.html}, abstract = {Understanding the inner representation of a neural network helps users improve models. Concept-based methods have become a popular choice for explaining deep neural networks post-hoc because, unlike most other explainable AI techniques, they can be used to test high-level visual "concepts" that are not directly related to feature attributes. For instance, the concept of "stripes" is important to classify an image as a zebra. Concept-based explanation methods, however, require practitioners to guess and manually collect multiple candidate concept image sets, making the process labor-intensive and prone to overlooking important concepts. Addressing this limitation, in this paper, we frame concept image set creation as an image generation problem. However, since naively using a standard generative model does not result in meaningful concepts, we devise a reinforcement learning-based preference optimization (RLPO) algorithm that fine-tunes a vision-language generative model from approximate textual descriptions of concepts. Through a series of experiments, we demonstrate our method’s ability to efficiently and reliably articulate diverse concepts that are otherwise challenging to craft manually.} }
Endnote
%0 Conference Paper %T Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks’ Internal Representations %A Aditya Taparia %A Som Sagar %A Ransalu Senanayake %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-taparia25a %I PMLR %P 59154--59181 %U https://proceedings.mlr.press/v267/taparia25a.html %V 267 %X Understanding the inner representation of a neural network helps users improve models. Concept-based methods have become a popular choice for explaining deep neural networks post-hoc because, unlike most other explainable AI techniques, they can be used to test high-level visual "concepts" that are not directly related to feature attributes. For instance, the concept of "stripes" is important to classify an image as a zebra. Concept-based explanation methods, however, require practitioners to guess and manually collect multiple candidate concept image sets, making the process labor-intensive and prone to overlooking important concepts. Addressing this limitation, in this paper, we frame concept image set creation as an image generation problem. However, since naively using a standard generative model does not result in meaningful concepts, we devise a reinforcement learning-based preference optimization (RLPO) algorithm that fine-tunes a vision-language generative model from approximate textual descriptions of concepts. Through a series of experiments, we demonstrate our method’s ability to efficiently and reliably articulate diverse concepts that are otherwise challenging to craft manually.
APA
Taparia, A., Sagar, S. & Senanayake, R.. (2025). Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks’ Internal Representations. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:59154-59181 Available from https://proceedings.mlr.press/v267/taparia25a.html.

Related Material