Zero-shot Concept Bottleneck Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yamaguchi, Shin'ya, Nishida, Kosuke, Chijiwa, Daiki, Ida, Yasutoshi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908933889720320
author Yamaguchi, Shin'ya
Nishida, Kosuke
Chijiwa, Daiki
Ida, Yasutoshi
author_facet Yamaguchi, Shin'ya
Nishida, Kosuke
Chijiwa, Daiki
Ida, Yasutoshi
contents Concept bottleneck models (CBMs) are inherently interpretable and intervenable neural network models, which explain their final label prediction by the intermediate prediction of high-level semantic concepts. However, they require target task training to learn input-to-concept and concept-to-label mappings, incurring target dataset collections and training resources. In this paper, we present zero-shot concept bottleneck models (Z-CBMs), which predict concepts and labels in a fully zero-shot manner without training neural networks. Z-CBMs utilize a large-scale concept bank, which is composed of millions of vocabulary extracted from the web, to describe arbitrary input in various domains. For the input-to-concept mapping, we introduce concept retrieval, which dynamically finds input-related concepts by the cross-modal search on the concept bank. In the concept-to-label inference, we apply concept regression to select essential concepts from the retrieved concepts by sparse linear regression. Through extensive experiments, we confirm that our Z-CBMs provide interpretable and intervenable concepts without any additional training. Code will be available at https://github.com/yshinya6/zcbm.
format Preprint
id arxiv_https___arxiv_org_abs_2502_09018
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Zero-shot Concept Bottleneck Models
Yamaguchi, Shin'ya
Nishida, Kosuke
Chijiwa, Daiki
Ida, Yasutoshi
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Concept bottleneck models (CBMs) are inherently interpretable and intervenable neural network models, which explain their final label prediction by the intermediate prediction of high-level semantic concepts. However, they require target task training to learn input-to-concept and concept-to-label mappings, incurring target dataset collections and training resources. In this paper, we present zero-shot concept bottleneck models (Z-CBMs), which predict concepts and labels in a fully zero-shot manner without training neural networks. Z-CBMs utilize a large-scale concept bank, which is composed of millions of vocabulary extracted from the web, to describe arbitrary input in various domains. For the input-to-concept mapping, we introduce concept retrieval, which dynamically finds input-related concepts by the cross-modal search on the concept bank. In the concept-to-label inference, we apply concept regression to select essential concepts from the retrieved concepts by sparse linear regression. Through extensive experiments, we confirm that our Z-CBMs provide interpretable and intervenable concepts without any additional training. Code will be available at https://github.com/yshinya6/zcbm.
title Zero-shot Concept Bottleneck Models
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.09018