Saved in:
Bibliographic Details
Main Authors: Fletcher, Luan, van der Klis, Robert, Sedláček, Martin, Vasilev, Stefan, Athanasiadis, Christos
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.13989
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910656148537344
author Fletcher, Luan
van der Klis, Robert
Sedláček, Martin
Vasilev, Stefan
Athanasiadis, Christos
author_facet Fletcher, Luan
van der Klis, Robert
Sedláček, Martin
Vasilev, Stefan
Athanasiadis, Christos
contents The growing reproducibility crisis in machine learning has brought forward a need for careful examination of research findings. This paper investigates the claims made by Lei et al. (2023) regarding their proposed method, LICO, for enhancing post-hoc interpretability techniques and improving image classification performance. LICO leverages natural language supervision from a vision-language model to enrich feature representations and guide the learning process. We conduct a comprehensive reproducibility study, employing (Wide) ResNets and established interpretability methods like Grad-CAM and RISE. We were mostly unable to reproduce the authors' results. In particular, we did not find that LICO consistently led to improved classification performance or improvements in quantitative and qualitative measures of interpretability. Thus, our findings highlight the importance of rigorous evaluation and transparent reporting in interpretability research.
format Preprint
id arxiv_https___arxiv_org_abs_2410_13989
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reproducibility study of "LICO: Explainable Models with Language-Image Consistency"
Fletcher, Luan
van der Klis, Robert
Sedláček, Martin
Vasilev, Stefan
Athanasiadis, Christos
Computer Vision and Pattern Recognition
The growing reproducibility crisis in machine learning has brought forward a need for careful examination of research findings. This paper investigates the claims made by Lei et al. (2023) regarding their proposed method, LICO, for enhancing post-hoc interpretability techniques and improving image classification performance. LICO leverages natural language supervision from a vision-language model to enrich feature representations and guide the learning process. We conduct a comprehensive reproducibility study, employing (Wide) ResNets and established interpretability methods like Grad-CAM and RISE. We were mostly unable to reproduce the authors' results. In particular, we did not find that LICO consistently led to improved classification performance or improvements in quantitative and qualitative measures of interpretability. Thus, our findings highlight the importance of rigorous evaluation and transparent reporting in interpretability research.
title Reproducibility study of "LICO: Explainable Models with Language-Image Consistency"
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.13989