{"id":26359015,"url":"https://github.com/zcccccz/causalwalk","last_synced_at":"2025-03-16T15:58:31.255Z","repository":{"id":212106507,"uuid":"730521027","full_name":"zcccccz/CausalWalk","owner":"zcccccz","description":"Code and data for our paper \"Causal Walk: Debiasing Multi-hop Fact Verification with Front-door Adjustment\" in AAAI 2024.","archived":false,"fork":false,"pushed_at":"2024-05-30T10:28:48.000Z","size":513,"stargazers_count":3,"open_issues_count":1,"forks_count":1,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-05-30T12:04:56.396Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"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/zcccccz.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}},"created_at":"2023-12-12T05:26:11.000Z","updated_at":"2024-05-30T10:28:52.000Z","dependencies_parsed_at":"2023-12-12T15:55:46.606Z","dependency_job_id":null,"html_url":"https://github.com/zcccccz/CausalWalk","commit_stats":null,"previous_names":["zcccccz/causalwalk"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zcccccz%2FCausalWalk","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zcccccz%2FCausalWalk/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zcccccz%2FCausalWalk/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zcccccz%2FCausalWalk/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/zcccccz","download_url":"https://codeload.github.com/zcccccz/CausalWalk/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":243893874,"owners_count":20364916,"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":[],"created_at":"2025-03-16T15:58:30.859Z","updated_at":"2025-03-16T15:58:31.247Z","avatar_url":"https://github.com/zcccccz.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Causal Walk\nCode and data for our paper \"[Causal Walk: Debiasing Multi-hop Fact Verification with Front-door Adjustment](https://doi.org/10.1609/aaai.v38i17.29925)\" in AAAI 2024.\n![CausalWalk](CausalWalk.svg)\n\n# Download\nTo reproduce the results in the paper, you need to download the dataset and our pre-trained model from [here](https://drive.google.com/drive/folders/19IdPsf0hZCnoHVnfHIelzEmnbkU3EF_6?usp=sharing).\n\nThen you should organize them in the following format.\n```\nCausalWalk\n    ├── data\n    ├── pretrained_models\n    ├── outputs\n    ├── data_load_utils.py\n    ├── models.py\n    ├── train_causal_walk_fever_2way.py\n    ├── train_causal_walk_politihop_2way.py\n    ├── train_causal_walk_politihop_3way.py\n    ├── train.sh\n    └── utils.py\n```\n\n# Environment\nInstall python dependencies.\n```\npip install -r requirements.txt\n```\n\n# Reproduction\nTrained on the FEVER dataset and evaluated on Adversarial FEVER. (2-way)\n```\nCUDA_VISIBLE_DEVICES=\"0\" python train_causal_walk_fever_2way.py \\\n--seed 1234 \\\n--batch_size 16 \\\n--lr 2e-5 \\\n--epochs 20 \\\n--weight_decay 5e-4 \\\n--evi_num 5 \\\n--max_seq_length 128\n```\nTrained on the PolitiHop dataset and evaluated on Adversarial PolitiHop. (3-way)\n```\nCUDA_VISIBLE_DEVICES=\"0\" python train_causal_walk_politihop_3way.py \\\n--seed 1234 \\\n--batch_size 4 \\\n--lr 1e-5 \\\n--epochs 20 \\\n--weight_decay 5e-4 \\\n--evi_num 20 \\\n--max_seq_length 128 \n```\nTrained on the PolitiHop dataset and evaluated on Symmetric PolitiHop. (2-way)\n```\nCUDA_VISIBLE_DEVICES=\"0\" python train_causal_walk_politihop_2way.py \\\n--seed 1234 \\\n--batch_size 4 \\\n--lr 1e-5 \\\n--epochs 10 \\\n--weight_decay 5e-4 \\\n--evi_num 20 \\\n--max_seq_length 128 \n```\n# Citation\n```\n@inproceedings{zhang2024causal,\n  title={Causal Walk: Debiasing Multi-Hop Fact Verification with Front-Door Adjustment},\n  author={Zhang, Congzhi and Zhang, Linhai and Zhou, Deyu},\n  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},\n  volume={38},\n  number={17},\n  pages={19533--19541},\n  year={2024}\n}\n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzcccccz%2Fcausalwalk","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzcccccz%2Fcausalwalk","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzcccccz%2Fcausalwalk/lists"}