Hermes: A Modular Multi-Agent System for StructuringClinical Text
Bibliographic record
Abstract
In today's age of information, unstructured information can become overwhelming and difficult to interpret, particularly in safety critical domains such as healthcare where the volume and complexity of unstructured textual notes is required to be interpretable, insightful, and easily automated for processing. This paper introduces Hermes, a modular agentic system that transforms unstructured clinical text into a modified version of the Subjective-Objective-Assessment-Plan (SOAP) format and generates a knowledge graph offering a high-level, distilled view that facilitates downstream clinical reasoning and decision-making. Hermes employs a multi-agent architecture consisting of four specialized components: Hermes-R (report generation), Hermes-G (knowledge graph generation), Hermes-Q (question-answer pair generation), and Hermes-A (answer generation). These agents operate sequentially with validation to generate structured medical information using iterative refinement. Preliminary evaluations on a few samples demonstrate that Hermes is able to generate structured clinical reports and knowledge graphs according to provided specifications from unstructured discharge summaries with good consistency, accuracy, and reward score. Hermes offers a unified framework that advances clinical natural language processing, bridging structured representation, question answering, and semantic validation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.010 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".