v287: Proceedings of CHIL 2025

2026-06-04 · Source: Proceedings of Machine Learning Research · Field: Health & Wellbeing — Clinical Care & Medical Practice, Medical Devices & Health Technology, Health & Medical Research · Depth: Expert, medium

Summary

Volume 287 presents the proceedings of the sixth Conference on Health, Inference, and Learning (CHIL), held from June 25-27, 2025, at UC Berkeley. This collection features 36 research papers exploring diverse applications of machine learning in healthcare. Key topics include developing disease progression models that address health disparities, uncovering knowledge gaps in radiology report generation using knowledge graphs, and integrating medical ontologies for robust clinical code embeddings. Other significant contributions cover benchmarking ECG delineation with deep neural networks, leveraging electrocardiogram-language models for few-shot question answering, and a case study on synthetic medical record generation with commercial LLMs. The volume also addresses multiaccuracy for subpopulation calibration, advancing sleep staging with smartwatch data, and causal inference in GWAS. Further research explores multimodal wearable sensor data for stress detection, predicting temporal changes in chest X-rays, and the utility of LLMs in understanding clinical text and extracting dense information from case reports.

Key takeaway

For AI Scientists and Machine Learning Engineers developing healthcare solutions, you should prioritize models that address health disparities and ensure subpopulation calibration. Integrate medical ontologies to enhance data robustness and explore multimodal approaches for richer insights from EHRs, imaging, and wearables. Your focus should extend to causal inference for genetic studies and robust evaluation frameworks for LLMs in clinical text, mitigating risks of bias and ensuring interpretability in real-world applications.

Key insights

The proceedings highlight diverse advancements in machine learning applications across healthcare domains.

Principles

• Disease progression models must capture health disparities.
• Integrating medical ontologies enhances clinical data robustness.
• Causal considerations improve machine learning utility in GWAS.

Method

Knowledge graphs can uncover gaps in radiology report generation models. Contrastive pretraining is effective for multimodal wearable sensor data. Test-time calibration personalizes biosignal adaptation.

In practice

• Utilize smartwatches for sleep staging and stress detection.
• Explore LLMs for synthetic medical record generation.
• Apply VR and robotics for upper extremity training.

Topics

• Clinical Machine Learning
• Medical Imaging
• Large Language Models
• Wearable Health Tech
• Causal Inference
• Health Equity

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• v298: Proceedings of MLHC 2025 — Advanced ML techniques are transforming healthcare diagnostics, prognostics, and treatment across diverse modalities.
• Mapping neighbourhood-level drivers of type 2 diabetes for precision public health using predictive and causal machine learning — Neighbourhood-level factors, identified via ML and causal inference, predict Type 2 diabetes risk and inform targeted public health.
• Calibre emerges from stealth with $3.3M to tackle “health guesswork” through causal AI — Causal AI combined with clinical expertise can deliver personalized, proactive health navigation at an accessible price point.
• v297: Proceedings of ML4H 2025 — Machine learning advancements across LLMs, medical imaging, and time-series analysis are transforming healthcare applications.
• Beyond Semantic Similarity: A Component-Wise Evaluation Framework for Medical Question Answering Systems with Health Equity Implications — Current LLM medical QA evaluations overlook critical accuracy and health equity risks.
• Bridging the interpretability gap for medical artificial intelligence models using class-association manifold learning — CAML enhances medical AI explainability by decoupling decision patterns and visualizing global class-associated knowledge.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Proceedings of Machine Learning Research.