Skip to main content
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arXiv logo
Cornell University Logo

Computer Science > Machine Learning

arXiv:2109.11160 (cs)
[Submitted on 23 Sep 2021 (v1), last revised 17 Feb 2022 (this version, v2)]

Title:Toward a Unified Framework for Debugging Concept-based Models

Authors:Andrea Bontempelli, Fausto Giunchiglia, Andrea Passerini, Stefano Teso
View PDF
Abstract:In this paper, we tackle interactive debugging of "gray-box" concept-based models (CBMs). These models learn task-relevant concepts appearing in the inputs and then compute a prediction by aggregating the concept activations. Our work stems from the observation that in CBMs both the concepts and the aggregation function can be affected by different kinds of bugs, and that fixing these bugs requires different kinds of corrective supervision. To this end, we introduce a simple schema for human supervisors to identify and prioritize bugs in both components, and discuss solution strategies and open problems. We also introduce a novel loss function for debugging the aggregation step that generalizes existing strategies for aligning black-box models to CBMs by making them robust to how the concepts change during training.
Comments: 11 pages, 1 figure. Accepted at the AAAI-22 Workshop on Interactive Machine Learning
Subjects: Machine Learning (cs.LG); Software Engineering (cs.SE)
Cite as: arXiv:2109.11160 [cs.LG]
  (or arXiv:2109.11160v2 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2109.11160

Submission history

From: Stefano Teso [view email]
[v1] Thu, 23 Sep 2021 06:12:17 UTC (299 KB)
[v2] Thu, 17 Feb 2022 11:58:01 UTC (38 KB)
Full-text links:

Access Paper:

  • View PDF
  • TeX Source
  • Other Formats
view license
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2021-09
Change to browse by:
cs
cs.SE

References & Citations

DBLP - CS Bibliography

listing | bibtex
Andrea Passerini
Stefano Teso
export BibTeX citation

Bookmark

BibSonomy logo Reddit logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
IArxiv Recommender (What is IArxiv?)

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)