Research map / Large language models / AI and society
PyHealth 2.0: A Comprehensive Open-Source Toolkit for Accessible and Reproducible Clinical Deep Learning
John Wu, Yongda Fan, Zhenbang Wu, Paul Landes, Eric Schrock, Sayeed Sajjad Razin, Arjun Chatterjee, Naveen Baskaran, Joshua Steier, Andrea Fitzpatrick, Bilal Arif, Rian Atri and 5 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| run PyHealth 2.0 on a MacBook instead of PyHealth 1.16 on the server workstation | labevents mortality task · MIMIC-IV mortality, 8 workers | wall time (s): 2,235 -> 1,357 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 | MIMIC-IV mortality · 16 workers peak memory | GB: 457.42 -> 28.70 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 | MIMIC-IV drug recommendation · 16 workers | wall time (s): 802 -> 569 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 | MIMIC-IV length of stay · 16 workers | wall time (s): 1,922 -> 1,096 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 | MIMIC-IV mortality · 16 workers wall time | s: 2,235 -> 2,385 | Hurt |
| replace PyHealth 1.16 with PyHealth 2.0 | MIMIC-IV mortality · 1 worker wall time | s: 6,841 -> 17,671 | Hurt |
| replace Pandas with PyHealth 2.0 multi-worker | MIMIC-IV mortality · 1 worker vs 16 workers | wall time (s): 93,708 -> 2,385 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 data loading | MIMIC-IV dataloading · load and cache time | s: 3548.90 -> 1093.38 | Improved |
| replace PyHealth 1.16 with PyHealth 2.0 patient access | MIMIC-IV random patient access · patient access time | ms: 0.01 -> 208.85 | Hurt |
| use PyHealth 2.0 instead of 1.16 for patient exploration | patient exploration · lines of code | lines: 14 -> 10 | Improved |
| add neighborhood conformal prediction over vanilla conformal prediction | TUEV conformal prediction · TUEV (alpha=0.1) | empirical coverage: 0.7461 -> 0.9152 | Improved |
| add KDE covariate shift adjustment to conformal prediction | TUEV conformal prediction · TUEV (alpha=0.1) | empirical coverage: 0.7461 -> 0.7457 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Pandas — PANDAS: Improving Many-shot Jailbreaking via Positive Affirmation, Negative Demonstration, and Adaptive Sampling (ICML 2025) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CLIMB: Data Foundations for Large Scale Multimodal Clinical Foundation Models (ICML 2025 · similarity 0.85)
- CARES: A Comprehensive Benchmark of Trustworthiness in Medical Vision Language Models (NeurIPS 2024 · similarity 0.85)
- Beyond Classification Accuracy: Neural-MedBench and the Need for Deeper Reasoning Benchmarks (ICLR 2026 · similarity 0.84)
- QoQ-Med: Building Multimodal Clinical Foundation Models with Domain-Aware GRPO Training (NeurIPS 2025 · similarity 0.84)
- PyTDC: A multimodal machine learning training, evaluation, and inference platform for biomedical foundation models (ICML 2025 · similarity 0.84)
- CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
