{"id":17819212,"url":"https://github.com/baldassarrefe/graph-network-explainability","last_synced_at":"2025-03-18T07:30:28.929Z","repository":{"id":71581115,"uuid":"175638717","full_name":"baldassarreFe/graph-network-explainability","owner":"baldassarreFe","description":"Explainability techniques for Graph Networks, applied to a synthetic dataset and an organic chemistry task. Code for the workshop paper \"Explainability Techniques for Graph Convolutional Networks\" (ICML19)","archived":false,"fork":false,"pushed_at":"2019-11-12T18:37:22.000Z","size":6791,"stargazers_count":122,"open_issues_count":3,"forks_count":16,"subscribers_count":7,"default_branch":"master","last_synced_at":"2025-02-28T07:51:59.617Z","etag":null,"topics":["artificial-intelligence","bioinformatics","explainability","graph-networks"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/baldassarreFe.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2019-03-14T14:30:38.000Z","updated_at":"2024-12-18T07:19:45.000Z","dependencies_parsed_at":"2023-04-04T00:00:40.504Z","dependency_job_id":null,"html_url":"https://github.com/baldassarreFe/graph-network-explainability","commit_stats":null,"previous_names":[],"tags_count":1,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/baldassarreFe%2Fgraph-network-explainability","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/baldassarreFe%2Fgraph-network-explainability/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/baldassarreFe%2Fgraph-network-explainability/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/baldassarreFe%2Fgraph-network-explainability/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/baldassarreFe","download_url":"https://codeload.github.com/baldassarreFe/graph-network-explainability/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":243910715,"owners_count":20367538,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["artificial-intelligence","bioinformatics","explainability","graph-networks"],"created_at":"2024-10-27T16:59:50.281Z","updated_at":"2025-03-18T07:30:28.922Z","avatar_url":"https://github.com/baldassarreFe.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Explainability Techniques for Graph Convolutional Networks\n\nCode and notebooks for the paper [\"Explainability Techniques for Graph Convolutional Networks\"](https://arxiv.org/abs/1905.13686) \naccepted at the ICML 2019 Workshop [\"Learning and Reasoning with Graph-Structured Data\"](https://graphreason.github.io/).\n\n## Overview\nA Graph Network trained to predict the solubility of organic molecules is applied to _sucrose_, \nthe prediction is explained using [Layer-wise Relevance Propagation](http://heatmapping.org) that assigns \npositive and negative relevance to the nodes and edges of the molecular graph: \n\n![Sucrose solubility LRP](resources/sucrose.png)\n\nThe predicted solubility can be broken down to the individual features of the atoms and their bonds:\n\n![Sucrose solubility LRP nodes](resources/sucrose-atoms.png)\n![Sucrose solubility LRP edges](resources/sucrose-bonds.png)\n\n## Code structure\n- `src`, `config`, `data` contain code, configuration files and data for the experiments\n  - `infection`, `solubility` contain the code for the two experiments in the paper\n  - `torchgraphs` contain the core graph network library\n  - `guidedbackrprop`, `relevance` contain the code to run Guided Backpropagation and Layer-wise Relevance Propagation on top of PyTorch's `autograd`\n- `notebooks`, `models` contain a visualization of the datasets, the trained models and the results of our experiments\n- `test` contains unit tests for the `torchgraphs` module (core GN library)\n- `conda.yaml` contains the conda environment for the project\n\n## Setup\nThe project is build on top of Python 3.7, PyTorch 1.1+, \n[torchgraphs](https://github.com/baldassarreFe/torchgraphs) 0.0.1 and many other open source projects.\n\nA [Conda](https://conda.io) environment for the project can be installed as: \n```bash\nconda env create -n gn-exp -f conda.yaml\nconda activate gn-exp\npython setup.py develop\npytest\n```\n\n## Training\nDetailed instructions for data processing, training and hyperparameter search can be found in the respective subfolders:  \n- Infection: [infection/notes.md](./src/infection/notes.md)\n- Solubility: [solubility/notes.md](./src/solubility/notes.md)\n\n## Experimental results\nThe results of our experiments are visualized through the notebooks in [`notebooks`](./notebooks):\n```bash\nconda activate gn-exp\ncd notebooks\njupyter lab \n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbaldassarrefe%2Fgraph-network-explainability","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbaldassarrefe%2Fgraph-network-explainability","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbaldassarrefe%2Fgraph-network-explainability/lists"}