HalluciGuard-eye: knowledge graph-enhanced retrieval-augmented generation for hallucination detection and mitigation in ophthalmic multimodal large language models.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: HalluciGuard-Eye combines knowledge-graph retrieval and hierarchical factuality scoring; across four ophthalmic imaging benchmarks it reports 42.2% lower hallucination rates and 91.6% factual accuracy on EyePACS, while diabetic-retinopathy sensitivity/specificity reached 89.6%/92.1%. This is a benchmark result, not clinical deployment evidence.
Technology: Knowledge-graph retrieval-augmented multimodal language model with factuality scoring
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Knowledge-graph RAG, visual-semantic alignment and factuality grounding reduce errors in multimodal ophthalmic LLM outputs.
- Limitation / still unknown
- Four public benchmarks, component ablation and reported metrics checked; benchmark validation does not establish prospective clinical safety.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2026.1976190/full
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.