[Model] Implemented SubgraphX Explainer for Homogeneous graph #5315
Conversation
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
@mufeili I chose the first homogeneous graph defined, which has 10 nodes and 17 edges. It was completed in ~3 mins. Can you trigger the
|
|
@kunmukh That's great! Is it possible to reduce the test time to say 20 seconds by using an alternative graph and tuning various parameters? Let's say the CI will time out after 1 hour and we might have at least 100 things to test. We will not want to spend too much time on testing SubgraphX alone. |
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
1 similar comment
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
|
@mufeili got it under 20 sec. I changed the parameters to focus less on exploration. Can you trigger the bot and merge the PR 💯 ? |
|
Not authorized to trigger CI. Please ask core developer to help trigger via issuing comment:
|
) * subgraphx commit * nits * newline eof added * lint fix * test script updated to use default values * lint fix * graphs that are used for test cases are updated to a small graph * lint formatted * test paramter adj to complete the test under 20s * lint fixes --------- Co-authored-by: kxm180046 <[email protected]>
) * subgraphx commit * nits * newline eof added * lint fix * test script updated to use default values * lint fix * graphs that are used for test cases are updated to a small graph * lint formatted * test paramter adj to complete the test under 20s * lint fixes --------- Co-authored-by: kxm180046 <[email protected]>
Description
Implemented SubgraphX Explainer (https://arxiv.org/pdf/2102.05152.pdf) for Homogeneous Graphs. Original PR: #4913, this PR was created as the original PR went out of date.
Checklist
Please feel free to remove inapplicable items for your PR.
Changes
The output of the model has been manually verified on MUTAG dataset. A visualization script was created to visualize the mutag dataset (external ref: https://gist.github.com/kunmukh/06d948b6b730447baa932873026e15da). Unfortunately, there's no ground truth for the explanation except for what is mentioned in the research paper, so we cannot quantify the model's performance. Instead, we manually investigated the results and confirmed that they aligned well with human intuition and findings of the research paper.
Inline documentation is added following the original style, and an example of using the module is included in the documentation.