{"id":22856696,"url":"https://github.com/aisuko/multimodal-mimic","last_synced_at":"2025-05-07T07:44:48.493Z","repository":{"id":267242141,"uuid":"868319528","full_name":"Aisuko/multimodal-mimic","owner":"Aisuko","description":"Multi-modal LLM and traditional ML models for ICU modality prediction on MIMIC-III across various time windows.","archived":false,"fork":false,"pushed_at":"2025-03-07T01:24:06.000Z","size":12823,"stargazers_count":1,"open_issues_count":1,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-03-07T01:26:37.378Z","etag":null,"topics":["ai","logistic-regression","mimic-iii","modality-classification","multi-modal","neural-network","random-forest-classifier","transfer-learning","xgboost-classifier"],"latest_commit_sha":null,"homepage":"","language":"Python","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/Aisuko.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":"CITATION.cff","codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-10-06T04:14:35.000Z","updated_at":"2025-03-07T01:16:09.000Z","dependencies_parsed_at":"2025-03-03T09:45:25.926Z","dependency_job_id":null,"html_url":"https://github.com/Aisuko/multimodal-mimic","commit_stats":null,"previous_names":["aisuko/multimodal-mimic"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Aisuko%2Fmultimodal-mimic","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Aisuko%2Fmultimodal-mimic/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Aisuko%2Fmultimodal-mimic/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Aisuko%2Fmultimodal-mimic/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Aisuko","download_url":"https://codeload.github.com/Aisuko/multimodal-mimic/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246436093,"owners_count":20776965,"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":["ai","logistic-regression","mimic-iii","modality-classification","multi-modal","neural-network","random-forest-classifier","transfer-learning","xgboost-classifier"],"created_at":"2024-12-13T08:09:59.988Z","updated_at":"2025-03-31T07:46:01.813Z","avatar_url":"https://github.com/Aisuko.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\r\n    \u003ch1 align=\"center\"\u003e\r\n        When Simpler Is Better: Traditional Models Outperform LLMs in ICU Mortality Prediction\r\n    \u003c/h1\u003e\r\n     \u003cp\u003eThis study compares traditional machine learning models to a multi-modal LLM-based model for predicting ICU mortality using the MIMIC-III dataset. We test several time windows (6, 12, 18, 24, and 48 hours) after admission. The results show that traditional models, especially Random Forest, consistently perform better and are more efficient than the LLM-based model. Our analysis finds that higher feature correlation, steady data patterns, and balanced variability lead to better predictions. While LLMs have potential, their current complexity and longer training times make them less practical without careful data selection and preparation. These findings highlight the importance of choosing both the right model and the right time windows to achieve reliable ICU mortality predictions.\u003c/p\u003e\r\n\u003c/p\u003e\r\n\r\n\u003cp align=\"center\"\u003e\r\n  \u003cimg src=\"./imgs/Figure1-accuracy-time-windows.png\" alt=\"\" width=\"30%\" style=\"display: inline-block; margin: 0 1%;\" /\u003e\r\n  \u003cimg src=\"./imgs/auc-roc.png\" alt=\"\" width=\"30%\" style=\"display: inline-block; margin: 0 1%;\" /\u003e\r\n  \u003cimg src=\"./imgs/f1.png\" alt=\"\" width=\"30%\" style=\"display: inline-block; margin: 0 1%;\" /\u003e\r\n\u003c/p\u003e\r\n\r\n\r\n# Dataset\r\n\r\nPlease check [document of dataset](./documents/dataset.md)\r\n\r\n\u003cp align=\"center\"\u003e\r\n  \u003cimg src=\"./imgs/result_of_evaluation_ds.png\" alt=\"\" width=\"50%\" style=\"display: inline-block; margin: 0 2%;\" /\u003e\r\n\u003c/p\u003e\r\n\r\n\r\n\r\n# Training\r\n\r\nWe utilize a customized development container (devcontainer) to conduct all experiments within an isolated environment. This approach ensures consistency across development setups and mitigates issues related to Python dependencies. \r\n\r\nDifferent models have different training strategies, please check below:\r\n\r\n\u003cp align=\"center\"\u003e\r\n  \u003cimg src=\"./imgs/training_time.png\" alt=\"\" width=\"50%\" style=\"display: inline-block; margin: 0 2%;\" /\u003e\r\n\u003c/p\u003e\r\n\r\n\r\n## Training Customized LLM\r\n\r\nFor training the customized LLM model. Please use `tmux`\r\n\r\n```\r\ntmux new -s session_name\r\ntmux ls\r\ntmux a -t session_name\r\ntime python experiments/measurement_notes/measurement_notes_llm.py \u003e train_log.txt 2\u003e\u00261\r\nControl+B D\r\n\r\ntail -f train_log.txt\r\n```\r\n\r\n## Training Traditional Models\r\n\r\nFor training the traditional ML model, please use [Makefile](./Makefile).\r\n\r\n\r\n# Citation\r\n\r\n```bibtex\r\n@software{Li_Multimodal-mimic_2024,\r\nauthor = {Li, Bowen},\r\ndoi = {\u003c\u003e},\r\nmonth = dec,\r\ntitle = {{Multimodal-mimic}},\r\nurl = {https://github.com/Aisuko/multimodal-mimic},\r\nversion = {1.0.0},\r\nyear = {2024}\r\n}\r\n```\r\n\r\n\r\n# Acknowledgements\r\n\r\n* [Ryan King etc al.](https://github.com/kingrc15/multimodal-clinical-pretraining)\r\n* [YerevaNN](https://github.com/YerevaNN/mimic3-benchmarks)\r\n\r\nThanks for your contribution.\r\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faisuko%2Fmultimodal-mimic","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Faisuko%2Fmultimodal-mimic","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faisuko%2Fmultimodal-mimic/lists"}