concept-bottleneck-models

Concept Bottleneck Models, ICML 2020

Generalizable AI predicts immunotherapy outcomes across cancers and treatments

Code for the paper: Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discovery. ECCV 2024.

[NeurIPS 24] A new training and evaluation framework for learning interpretable deep vision models and benchmarking different interpretable concept-bottleneck-models (CBMs)

[ICLR 2025 Spotlight] This is the official repository for our paper: ''Enhancing Pre-trained Representation Classifiability can Boost its Interpretability''.

[CVPR 2025] Concept Bottleneck Autoencoder (CB-AE) -- efficiently transform any pretrained (black-box) image generative model into an interpretable generative concept bottleneck model (CBM) with minimal concept supervision, while preserving image quality

Code for the paper "CBVLM: Training-free Explainable Concept-based Large Vision Language Models for Medical Image Classification", Computers in Biology and Medicine, 2025.

Concept bottleneck models for multiview data with incomplete concept sets

[CVPR 2026] Explaining CLIP Zero-shot Predictions Through Concepts (EZPC)

Web Companion for Generalizable AI predicts immunotherapy outcomes across cancers and treatments

[MICCAI 2024] AdaCBM: An Adaptive Concept Bottleneck Model for Explainable and Accurate Diagnosis

Papers on CBMs with short descriptions of paper's content

Official code for the paper "Selective Concept Bottleneck Models Without Predefined Concepts" (TMLR 2025)

[CVPR'26 Findings] Vision-Language Models Encode Clinical Guidelines for Concept-Based Medical Reasoning

Code implementation for our paper "Semi-supervised Concept Bottleneck Models" (ICCV-2025)

[TMLR, 2025] This is the repository for: Revisiting Discover-then-Name Concept Bottleneck Models: A Reproducibility Study.

This project poses a new methodology for assessing and improving sequential concept bottleneck models (CBMs). The research undertaken in this project builds upon the model proposed by Grange et al., of which I was one of the co-authors.

Implementation of "Enhancing Concept Localization in CLIP-based Concept Bottleneck Models"

CBMJEV: Adaptive concept bottlenecks with fallible measurements. Research code, reproducibility tools, and editable figures.

PyTorch study of accuracy, interpretability and steerability in Concept Bottleneck Models on Fashion-MNIST.