{"id":18560863,"url":"https://github.com/harry24k/fgsm-pytorch","last_synced_at":"2025-04-10T02:31:12.878Z","repository":{"id":104396916,"uuid":"175375142","full_name":"Harry24k/FGSM-pytorch","owner":"Harry24k","description":"A pytorch implementation of \"Explaining and harnessing adversarial examples\"","archived":false,"fork":false,"pushed_at":"2019-09-04T14:46:50.000Z","size":715,"stargazers_count":67,"open_issues_count":1,"forks_count":16,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-03-24T15:42:03.643Z","etag":null,"topics":["adversarial-attacks","deep-learning","pytorch"],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","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/Harry24k.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","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":"2019-03-13T08:11:17.000Z","updated_at":"2025-03-11T05:06:03.000Z","dependencies_parsed_at":null,"dependency_job_id":"e91db93e-3a20-4c89-ace6-d32c51218d8c","html_url":"https://github.com/Harry24k/FGSM-pytorch","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Harry24k%2FFGSM-pytorch","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Harry24k%2FFGSM-pytorch/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Harry24k%2FFGSM-pytorch/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Harry24k%2FFGSM-pytorch/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Harry24k","download_url":"https://codeload.github.com/Harry24k/FGSM-pytorch/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248144193,"owners_count":21054881,"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":["adversarial-attacks","deep-learning","pytorch"],"created_at":"2024-11-06T22:04:59.079Z","updated_at":"2025-04-10T02:31:12.872Z","avatar_url":"https://github.com/Harry24k.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# FGSM-pytorch\n**A pytorch implementation of \"[Explaining and harnessing adversarial examples](https://arxiv.org/abs/1412.6572)\"**\n\n## Summary\nThis code is a pytorch implementation of **FGSM(Fast Gradient Sign Method).**   \nIn this code, I used FGSM to fool [Inception v3](https://arxiv.org/abs/1512.00567).   \nThe picture '[Giant Panda](http://www.image-net.org/)' is exactly the same as in the paper.   \nYou can add other pictures with a folder with the label name in the 'data'.   \n\n## Requirements\n* python==3.6   \n* numpy==1.14.2   \n* pytorch==1.0.0   \n\n## Important results not in the code\n- Mathmatical Results\n   - There are some important difference between adversarial training and L1 weight decay. (p.4)\n      - On logistic regression,\n      - Adversarial training : the L1 penalty is subtracted off inside of the activation during training.\n      - L1 weight decay : the L1 penalty is added to the training cost(=outside of the activation) during training.\n- Experimental Results\n   - We can use FGSM for a regularizer but it does not defend against all adversarial attack images. (p.5)\n   - RBF networks are resistant to adversarial examples, but not for Linear. (p.7)\n      - The author claims current methodologies all resemble the linear classifier, which is why do adversarial examples generalize\n   - Alternative hypotheses(generative models with input distribution, ensembles) are not resistant to adversarial examples. (p.8)\n\n## Notice\n- This Repository won't be updated.\n- Please check [the package of adversarial attacks in pytorch](https://github.com/Harry24k/adversairal-attacks-pytorch)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fharry24k%2Ffgsm-pytorch","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fharry24k%2Ffgsm-pytorch","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fharry24k%2Ffgsm-pytorch/lists"}