Skip to content

Latest commit

 

History

History
89 lines (60 loc) · 3.04 KB

README.md

File metadata and controls

89 lines (60 loc) · 3.04 KB
Raw

Landmark Explanation

Landmark Explanation is an Entity Matching specific explanation framework based on LIME.

The state of the art approaches for performing Entity Matching (EM) rely on machine & deep learning models for inferring pairs of matching / non-matching entities. Although the experimental evaluations demonstrate that these approaches are effective, their adoption in real scenarios is limited by the fact that they are difficult to interpret. Explainable AI systems have been recently proposed for complementing deep learning approaches. Their application to the scenario offered by EM is still new and requires to address the specificity of this task, characterized by particular dataset schemas, describing a pair of entities, and imbalanced classes. This paper introduces Landmark, a generic and extensible framework that extends the capabilities of a post-hoc perturbation-based explainer over the EM scenario. Landmark generates perturbations that take advantage of the particular schemas of the EM datasets, thus generating explanations more accurate and more interesting for the users than the ones generated by competing approaches.

Framework

Landmark flow

Quick Start: Landmark Explanation in 30 seconds

Run:

git clone https://github.com/softlab-unimore/landmark.git
pip install -r landmark/requirements.txt

Import Landmark

from landmark import Landmark

Initialize it passing the predict method of your model.

explainer = Landmark(model.predict, df, exclude_attrs=['id','label'], lprefix='left_', rprefix='right_')

Explain EM model over an element of your data.

el = df.sample()
exp = explainer.explain(el)

Plot your explanation

explainer.plot(exp, el)

Working example

Get started with Landmark Explanation!

Here you can find a working notebook in which we explain a Magellan model.

Here you can find a working notebook in which we explain a DeepMatcher model.

Reference

For a detailed description of the work please read our papers. Please cite the paper if you use the code from this repository in your work. Using landmarks for explaining entity matching models

@inproceedings{DBLP:conf/edbt/BaraldiBP021,
  author    = {Andrea Baraldi and
               Francesco Del Buono and
               Matteo Paganelli and
               Francesco Guerra},
  title     = {Using Landmarks for Explaining Entity Matching Models},
  booktitle = {{EDBT}},
  pages     = {451--456},
  publisher = {OpenProceedings.org},
  year      = {2021}
}

Check out our demo as well: Landmark explanation: An explainer for entity matching models

License

MIT License