RAG

Rosetta uses semantic retrieval to pull supporting material from notes, uploads, and medical databases.

Architecture

Text is embedded into vectors so retrieval can match meaning, not only keywords. Rosetta returns the closest passages from your available sources.

Semantic vs Keyword

Keyword Search (Traditional):

Query: "heart failure treatment"
Matches: Exact text "heart failure" AND "treatment"
Misses: "HFrEF management", "cardiac dysfunction therapy"

Semantic Search (Rosetta):

Query: "heart failure treatment"
Matches: Any semantically similar concepts
  - "HFrEF management"
  - "cardiac dysfunction therapy"
  - "GDMT for reduced EF"

Source Types

Rosetta organizes knowledge into two scopes:

1. Local: Sources specific to the current note and patient context.
2. Account: Guidelines, protocols, and papers in your library, available across notes.

PubMed Integration

Rosetta integrates with PubMed for evidence retrieval.

• Search: Converts a request into PubMed search terms.
• Retrieval: Fetches abstracts and metadata for top results.
• Ingestion: Processes results into the RAG store for citation use.

Usage Examples

Evidence-Based Treatment (HFrEF)

1. Add a “GDMT” PDF to the Account library.
2. In a note, ask: “heart failure reduced ejection fraction treatment”.
3. The system retrieves guidance on ACE-I/ARB, beta-blockers, and SGLT2i use.
4. The agent generates a plan with citations from the uploaded PDF.

Literature Review (AFib)

1. The agent searches PubMed for “atrial fibrillation anticoagulation 2024”.
2. It retrieves recent abstracts.
3. The agent generates a summary with current scoring criteria.

Institutional Protocols

1. Upload a hospital sepsis protocol to the Account library.
2. Query: “UCSF sepsis bundle timing”.
3. The agent returns the timing requirements from your protocol.

Next: Workflows | API