What this category's data looks like
This page addresses the research achievement registration and intellectual property management scenario. Materials include achievement registration forms, patent disclosure documents, and software copyright materials. Achievement registration forms are filled in by project teams at milestones, while patent disclosure documents and software copyright materials are prepared by inventors or agents. Once these three enter the knowledge base, they serve as retrieval corpus to support verification of fields, dates, and signatures during deployment, operations, and application handover. The update cadence follows registration and filing milestones, and versions change with material revisions. Field characteristics center on names, quantities, dates, signatures, and receipts. The same record is often scattered across multiple materials, so original text fragments must be retained during retrieval for later comparison.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate fragments and similarity scores (after pattern, merge, re-rank, filter, and upper-limit trimming). The handover-ready deployment status and operation records are completed by subsequent nodes or manually. Retrieval is only responsible for providing candidates and cited content; it does not generate statistical tables or authorization results.
| Object | Check | Verification action |
|---|---|---|
| Achievement registration form | Whether names, quantities, dates, and signatures correspond item by item with the patent disclosure document | Switch searchMode in the search test and compare whether semantic and full-text candidates both cover the two materials |
| Patent disclosure document | Whether versions are synchronized and whether dates and receipts are missing items | Compare the top few rankings before and after enabling usingReRank, and check the version and date fields in candidate fragments |
| Software copyright materials | Whether names, quantities, and dates for the same record do not match across the two materials | Record the number of candidates and the similarity score, and separately record the citation array length and the fragments after limit trimming |
| Deployment status and operation records | Whether version, connection, and task records are consistent | Check item by item against the handover sheet, see which record is missing, and confirm whether candidate fragments retain key original paragraphs |
Three common errors
- Symptom: Names, quantities, and dates for the same record do not match between the achievement registration form and the patent disclosure document. Cause: The two materials are filled in separately, and field alignment was not performed before retrieval.
- Symptom: Software copyright material versions are not synchronized, and old and new versions are recalled at the same time. Cause: The knowledge base corpus was not updated after material revision.
- Symptom: There are many candidates, but the top-ranked ones are irrelevant, and the answer lacks key paragraphs. Cause: The retrieval mode does not match the material form, or re-rank secondary sorting was not enabled.
How to verify the setup
This feature acceptance: Test once each with synonym-rewritten questions, compare candidates and similarity scores, then toggle usingReRank to observe the top few rankings, and check the citation array length under searchMode and limit.
Business test: Check item by item against the handover sheet once, see which record is missing, and confirm that version, connection, and task records are consistent.
- In the knowledge base search test, switch retrieval modes and compare whether candidate results include both semantic and full-text sources.
- Compare the order of the top few results before and after enabling re-rank, and separately record the citation array length and the tokens of the cited text.