Skip to content

Portable Patient Embeddings via LLM Summarization

Home / Concepts / Portable Patient Embeddings via LLM Summarization

Portable Patient Embeddings via LLM Summarization

Auto-generated stub. Edit this file to add more details.

Generating fixed-length, transferrable patient vector representations by first summarizing irregular ICU time series data into natural language using a frozen LLM, and then embedding the text summary.

Why It Matters

This is the core novel technique proposed to address the portability problem in clinical ML deployment.

Evidence

we map from irregular ICU time series onto concise natural language summaries using a frozen LLM, then embed each summary with a frozen text embedding model to obtain a fixed length vector

Metadata & Links