{"id":13481124,"url":"https://github.com/rctzeng/ICML19-EgoCNN","last_synced_at":"2025-03-27T11:31:50.601Z","repository":{"id":85996773,"uuid":"174918519","full_name":"rctzeng/ICML19-EgoCNN","owner":"rctzeng","description":"Code for \"Distributed, Egocentric Representations of Graphs for Detecting Critical Structures\" (ICML 2019)","archived":false,"fork":false,"pushed_at":"2021-08-24T08:06:03.000Z","size":168,"stargazers_count":19,"open_issues_count":0,"forks_count":4,"subscribers_count":1,"default_branch":"master","last_synced_at":"2024-10-30T14:43:32.240Z","etag":null,"topics":["convolutional-neural-networks","graph-convolutional-neural-networks","graph-embeddings","graph-networks","graph-neural-networks","icml","icml-2019","interpretability","scale-free-networks","self-similarity"],"latest_commit_sha":null,"homepage":"http://proceedings.mlr.press/v97/tzeng19a.html","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/rctzeng.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":"CITATION.cff","codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null}},"created_at":"2019-03-11T03:22:10.000Z","updated_at":"2024-06-16T10:58:32.000Z","dependencies_parsed_at":"2023-03-13T08:07:10.941Z","dependency_job_id":null,"html_url":"https://github.com/rctzeng/ICML19-EgoCNN","commit_stats":null,"previous_names":["rctzeng/icml19-egocnn"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rctzeng%2FICML19-EgoCNN","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rctzeng%2FICML19-EgoCNN/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rctzeng%2FICML19-EgoCNN/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rctzeng%2FICML19-EgoCNN/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/rctzeng","download_url":"https://codeload.github.com/rctzeng/ICML19-EgoCNN/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245836241,"owners_count":20680339,"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":["convolutional-neural-networks","graph-convolutional-neural-networks","graph-embeddings","graph-networks","graph-neural-networks","icml","icml-2019","interpretability","scale-free-networks","self-similarity"],"created_at":"2024-07-31T17:00:48.920Z","updated_at":"2025-03-27T11:31:50.306Z","avatar_url":"https://github.com/rctzeng.png","language":"Python","funding_links":[],"categories":["Deep Learning"],"sub_categories":[],"readme":"# Ego-CNN\nThis is the repo for \"Distributed, Egocentric Representations of Graphs for Detecting Critical Structures\", Ruo-Chun Tzeng, Shan-Hung Wu, In Proceedings of ICML 2019.\n * [slides](https://drive.google.com/open?id=1ypDgm_EVsJCjC0c5Rl7dBvzQSRjSbkXZ)\n\nIn the paper, we proposed **Ego-Convolution** layer, which keeps the nice properties of Convolution layer to the graph including:\n * detection of location-invariant patterns\n * enlarged receptive fields in multi-layer architecture\n * [most importantly] detection of **precise** patterns\n\nThis enables our Ego-CNN to provide explanation to its prediction when jointly learned with a task.\n![picture](figs/model-EgoCNN.png)\n 1. In effect, Ego-CNN with L layers can detect patterns up-to L-hop ego-networks.\n 2. By using the existing CNN visualization techniques such as Transposed Convolution or Grad-CAM variants, we can visualize the detected patterns in a specific filter or a specific neuron.\n 3. By tying the weight of filter across different layers, our Ego-CNN is regularized to detect **self-similar** patterns\n\n## Dependence\n * Python \u003e= 3.6\n * Tensorflow \u003e= 1.0\n * NetworkX 2.0\n * Numpy \u003e= 1.13, Matplotlib \u003e= 2.1\n * Optparse\n\n## To Reproduce Our Result On ICML'19\n\n### Step 1. Download and Preprocess Graph Classification Datasets\nExecute Command `python download_dataset.py` to download all the bioinformatic and social network datasets used in the paper.\n\n## Step 2. Train Ego-CNN on specified datasets for specified tasks\nTo reproduce ...\n * Graph Classification Experiments: run `./execute-graph-classification-on-benchmarks.sh`\n * Effectiveness of Scale-Free Regularizer: run `./execute-graph-classification-on-benchmarks.sh`\n * Visualization on synthetic compounds: run `./execute-graph-classification-on-benchmarks.sh`\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frctzeng%2FICML19-EgoCNN","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Frctzeng%2FICML19-EgoCNN","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frctzeng%2FICML19-EgoCNN/lists"}