{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/hirid-icu-benchmark-a-comprehensive-machine","title":"HiRID-ICU-Benchmark -- A Comprehensive Machine Learning Benchmark on High-resolution ICU Data","arxiv_id":"2111.08536","date":"2021-11-16","proceeding":"NeurIPS Datasets and Benchmarks 2021","authors":["Hugo Yèche","Rita Kuznetsova","Marc Zimmermann","Matthias Hüser","Xinrui Lyu","Martin Faltys","Gunnar Rätsch"],"abstract":"The recent success of machine learning methods applied to time series collected from Intensive Care Units (ICU) exposes the lack of standardized machine learning benchmarks for developing and comparing such methods. While raw datasets, such as MIMIC-IV or eICU, can be freely accessed on Physionet, the choice of tasks and pre-processing is often chosen ad-hoc for each publication, limiting comparability across publications. In this work, we aim to improve this situation by providing a benchmark covering a large spectrum of ICU-related tasks. Using the HiRID dataset, we define multiple clinically relevant tasks in collaboration with clinicians. In addition, we provide a reproducible end-to-end pipeline to construct both data and labels. Finally, we provide an in-depth analysis of current state-of-the-art sequence modeling methods, highlighting some limitations of deep learning approaches for this type of data. With this benchmark, we hope to give the research community the possibility of a fair comparison of their work.","url_abs":"https://arxiv.org/abs/2111.08536v4","url_pdf":"https://arxiv.org/pdf/2111.08536v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"hirid-icu-benchmark-a-comprehensive-machine","repo_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"circulatory-failure","task_name":"Circulatory Failure"},{"task_slug":"icu-mortality","task_name":"ICU Mortality"},{"task_slug":"kidney-function","task_name":"Kidney Function"},{"task_slug":"patient-phenotyping","task_name":"Patient Phenotyping"},{"task_slug":"remaining-length-of-stay","task_name":"Remaining Length of Stay"},{"task_slug":"respiratory-failure","task_name":"Respiratory Failure"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"LGBM","rank_in_archive_order":2,"of":8,"metrics":{"AUPRC":"0.389±0.003"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":3,"of":8,"metrics":{"AUPRC":"0.388±0.002"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"GRU","rank_in_archive_order":4,"of":8,"metrics":{"AUPRC":"0.368±0.005"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"Transformer","rank_in_archive_order":5,"of":8,"metrics":{"AUPRC":"0.352±0.006"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"LR","rank_in_archive_order":6,"of":8,"metrics":{"AUPRC":"0.305±0.000"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"LSTM","rank_in_archive_order":7,"of":8,"metrics":{"AUPRC":"0.32.2±0.008"},"uses_additional_data":false},{"leaderboard":"/sota/circulatory-failure-on-hirid","task":"Circulatory Failure","dataset":"HiRID","model":"TCN","rank_in_archive_order":8,"of":8,"metrics":{"AUPRC":"0.35.8±0.006"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":1,"of":7,"metrics":{"AUPRC":"0.626±0.000"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"Transformer","rank_in_archive_order":2,"of":7,"metrics":{"AUPRC":"0.610±0.008"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"GRU","rank_in_archive_order":3,"of":7,"metrics":{"AUPRC":"0.603 ±0.016"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"TCN","rank_in_archive_order":4,"of":7,"metrics":{"AUPRC":"0.602±0.011"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"LSTM","rank_in_archive_order":5,"of":7,"metrics":{"AUPRC":"0.600±0.009"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"Logistic Regression","rank_in_archive_order":6,"of":7,"metrics":{"AUPRC":"0.581±0.000"},"uses_additional_data":false},{"leaderboard":"/sota/icu-mortality-on-hirid","task":"ICU Mortality","dataset":"HiRID","model":"LGBM","rank_in_archive_order":7,"of":7,"metrics":{"AUPRC":"0.546±0.008"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":1,"of":6,"metrics":{"MAE":"0.45±0.00"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"LGBM","rank_in_archive_order":2,"of":6,"metrics":{"MAE":"0.45±0.00"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"Transformer","rank_in_archive_order":3,"of":6,"metrics":{"MAE":"0.48±0.02"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"GRU","rank_in_archive_order":4,"of":6,"metrics":{"MAE":"0.49±0.02"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"LSTM","rank_in_archive_order":5,"of":6,"metrics":{"MAE":"0.50±0.01"},"uses_additional_data":false},{"leaderboard":"/sota/kidney-function-on-hirid","task":"Kidney Function","dataset":"HiRID","model":"TCN","rank_in_archive_order":6,"of":6,"metrics":{"MAE":"0.50±0.01"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":1,"of":7,"metrics":{"Balanced Accuracy":"45.8±2.0"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"Transformer","rank_in_archive_order":2,"of":7,"metrics":{"Balanced Accuracy":"42.7±1.4"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"TCN","rank_in_archive_order":3,"of":7,"metrics":{"Balanced Accuracy":"41.6±2.3"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"LGBM","rank_in_archive_order":4,"of":7,"metrics":{"Balanced Accuracy":"40.4±0.8"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"LSTM","rank_in_archive_order":5,"of":7,"metrics":{"Balanced Accuracy":"39.5±1.2"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"GRU","rank_in_archive_order":6,"of":7,"metrics":{"Balanced Accuracy":"39.2±2.1"},"uses_additional_data":false},{"leaderboard":"/sota/patient-phenotyping-on-hirid","task":"Patient Phenotyping","dataset":"HiRID","model":"Logistic Regression","rank_in_archive_order":7,"of":7,"metrics":{"Balanced Accuracy":"39.1±0.0"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"LGBM","rank_in_archive_order":1,"of":6,"metrics":{"MAE":"56.9±0.4"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":2,"of":6,"metrics":{"MAE":"57.0±0.3"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"Transformer","rank_in_archive_order":3,"of":6,"metrics":{"MAE":"59.5±2.8"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"TCN","rank_in_archive_order":4,"of":6,"metrics":{"MAE":"59.8±2.8"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"GRU","rank_in_archive_order":5,"of":6,"metrics":{"MAE":"60.6±0.9"},"uses_additional_data":false},{"leaderboard":"/sota/remaining-length-of-stay-on-hirid","task":"Remaining Length of Stay","dataset":"HiRID","model":"LSTM","rank_in_archive_order":6,"of":6,"metrics":{"MAE":"60.7±1.6"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"LGBM ( + hand crafted features)","rank_in_archive_order":2,"of":8,"metrics":{"AUPRC":"0.604±0.002"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"Transformer","rank_in_archive_order":3,"of":8,"metrics":{"AUPRC":"0.594±0.003"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"GRU","rank_in_archive_order":4,"of":8,"metrics":{"AUPRC":"0.592±0.003"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"TCN","rank_in_archive_order":5,"of":8,"metrics":{"AUPRC":"0.589±0.003"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"LGBM","rank_in_archive_order":6,"of":8,"metrics":{"AUPRC":"0.585±0.001"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"LSTM","rank_in_archive_order":7,"of":8,"metrics":{"AUPRC":"0.569±0.003"},"uses_additional_data":false},{"leaderboard":"/sota/respiratory-failure-on-hirid","task":"Respiratory Failure","dataset":"HiRID","model":"Logistic Regression","rank_in_archive_order":8,"of":8,"metrics":{"AUPRC":"0.530±0.000"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2111.08536","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2111.08536"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/ratschlab/HIRID-ICU-Benchmark","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":4,"unverified":4},"by_repo_kind":{"official":{"samples":8,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"8c712712d170d6fa","entry":"balanced_accuracy_compute_fn","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/models/metrics.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/models/metrics.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8c712712d170d6fa"}},{"code_sha256_prefix":"98ad0404a78f5d45","entry":"convert_to_days","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/common/reference_data.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/common/reference_data.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"98ad0404a78f5d45"}},{"code_sha256_prefix":"1bdf730ebc38566a","entry":"ece_curve_compute_fn","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/models/metrics.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/models/metrics.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1bdf730ebc38566a"}},{"code_sha256_prefix":"1ecd73f1856cbb81","entry":"read_var_ref_table","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/common/lookups.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/common/lookups.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1ecd73f1856cbb81"}},{"code_sha256_prefix":"913d82af53065418","entry":"accuracy","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/models/metrics.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/models/metrics.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"913d82af53065418"}},{"code_sha256_prefix":"0af839c625fc0b73","entry":"get_bindings_and_params","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/models/utils.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/models/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0af839c625fc0b73"}},{"code_sha256_prefix":"7bad4865010364c1","entry":"is_df_sorted","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/labels/label_benchmark.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/labels/label_benchmark.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7bad4865010364c1"}},{"code_sha256_prefix":"b37afd4624d8fae8","entry":"read_reference_table","repo":"ratschlab/HIRID-ICU-Benchmark","repo_kind":"official","path":"icu_benchmarks/common/lookups.py","file_url":"https://github.com/ratschlab/HIRID-ICU-Benchmark/blob/HEAD/icu_benchmarks/common/lookups.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b37afd4624d8fae8"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}