LLMs serve as synthetic clinical experts for rare disease modeling

Researchers have developed a novel method using large language models (LLMs) as synthetic clinical experts to improve the modeling of rare diseases. This approach integrates clinical knowledge into the model fitting process by querying LLMs offline on patient observation descriptions to obtain expert judgments. These judgments are then used to train a differentiable surrogate model, augmenting the loss function of a variational autoencoder to ensure reconstructed data maintains clinical relevance. In a study involving spinal muscular atrophy, this technique reduced disagreement in disease type labels and enhanced the prediction of motor function milestones compared to unsupervised methods. AI

IMPACT LLMs can now be leveraged to inject clinical expertise into rare disease modeling, potentially improving diagnostic accuracy and predictive capabilities.

RANK_REASON Research paper detailing a novel application of LLMs in medical data modeling. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• Clemens Schächter
• DagsHub
• Gotit.pub
• Hugging Face
• large language models
• spinal muscular atrophy
• variational autoencoder

• paper
• model release
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →