SkinMind: Equipping Agentic RAG with Human-Inspired Graph Probing for Reliable Dermatology AI

Alex: Better medical AI may need better questions, not just more reference material. A dermatology system called SkinMind improved benchmark answers by choosing which linked evidence to inspect next. Its advantage wasn't simply reading more.

Sam: That sounds useful, but better benchmark answers aren't the same as safer medical advice. What kind of evidence are we discussing?

Alex: It's a method paper and arXiv preprint from Jing Hao and colleagues, including researchers at Ant Group. They introduce SkinMind and a dermatology evidence resource, then compare answering systems on public tests. The claim concerns evidence-grounded reasoning, not demonstrated safety in patient care.

Sam: So this matters for researchers deciding how to connect medical models to outside knowledge, rather than choosing a clinical diagnostic tool.

Alex: That's the audience. We'll unpack what the model controls, why the comparisons are informative, and where its search still goes wrong. The central tension is that finding relevant evidence doesn't guarantee asking the question that separates competing diagnoses.

Sam: Start with the existing approach. When a model already has access to medical references, what does SkinMind change?

Alex: Retrieval-augmented generation means fetching outside evidence to help produce an answer. Conventional versions search once, collect passages, and hand them to the model. But the initial question may not reveal everything the model will need as its reasoning develops.

Sam: Some systems already search repeatedly. Is the novelty repeated searching, or something about what they search?

Alex: It's the combination. Agentic retrieval lets a model decide when and how to search again. A knowledge graph represents concepts connected by named relationships. SkinMind lets the model explore that graph interactively, rather than merely receiving results selected from it.

Sam: So the authors are joining two existing directions. What distinction do they draw from graph-based methods such as LightRAG?

Alex: They argue that many graph-based systems mainly use the graph as a retrieval index. Selection follows similarity scores or predefined expansion rules. SkinMind instead exposes navigation choices: inspect a disease, list its relationships, follow a promising connection, or open the supporting source.

Sam: A disease could have a huge number of connections. How does the model avoid turning that into another oversized reading pile?

Alex: Through progressive disclosure: showing details only when needed. The model initially receives compact results, such as concept names and types. It can inspect available relationship types before selecting particular neighbours. Detailed descriptions and original passages require further requests.

Sam: That separates deciding where to look from reading everything there. What does the underlying resource contain?

Alex: The authors build a graph with about a million nodes. It draws on dermatology guidelines and textbooks, alongside clinical images. Extracted textual claims retain links to supporting passages, so the model can return from a summary to its source context.

Sam: Those connections are themselves machine-generated. How do the authors handle contradictory information?

Alex: Their construction process distinguishes genuine contradictions from claims about different clinical circumstances. For detected conflicts, a model searches for additional evidence and attempts resolution. Ambiguous cases go to experts. That is a documented resolution process, not proof that every graph statement is correct.

Sam: Make the navigation concrete. What happens in an actual diagnostic example?

Alex: In the paper's successful example, the model must identify a photographed skin condition from answer options. It first retrieves similar images and inspects their metadata. One result has a mismatch between its caption and aligned disease label, so that result isn't straightforward support for a diagnosis.

Sam: If a retrieved image has questionable metadata, what other evidence does the system pursue?

Alex: It reads descriptions of lichen amyloidosis and xerotic eczema, competing diagnoses in the question. Then it follows morphology relationships—connections describing lesion shape and appearance. Retrieved features include clustered papules, meaning small raised bumps, and a rough, cobblestone appearance.

Sam: So the useful move isn't just finding a similar photograph. It's checking whether the disease's described features fit the observed lesions.

Alex: That trajectory reaches lichen amyloidosis, the benchmark's correct answer. Direct answering and LightRAG select other options in this example. It's an illustration of the mechanism, though, not the main evidence that the method works.

Sam: Then let's move to the comparison. How do they separate better navigation from simply giving SkinMind better references?

Alex: They test six dermatology visual-question-answering benchmarks, meaning questions involving images. All retrieval methods use the same knowledge resource and shared retrieval backend. Within each answering-model setting, they also share task instructions and generation settings.

Sam: And they don't rely on a single answering model to make the approach look good?

Alex: They test both a GPT mini model and a Gemini Flash model. SkinMind improves the average score over direct answering by roughly nine points in each setting. It also beats the competing retrieval methods across the tested benchmarks.

Sam: Direct answering leaves outside knowledge unused. How much of the advantage survives against a stronger retrieval competitor?

Alex: Against LightRAG, the average gain is about six points with the GPT model. The advantage also holds with Gemini, although it's smaller. That supports the authors' argument that access to knowledge isn't enough; how the system explores it matters.

Sam: Medical image collections can overlap. Did they check whether retrieval was effectively supplying the test photograph?

Alex: They remove exact image duplicates and filter potential near-duplicates using image similarity. Those rules apply across the retrieval methods. Most benchmarks use multiple-choice accuracy; the free-response benchmark uses a model judge comparing answers with references, without seeing the input image.

Sam: That makes the comparison more credible within these tests. But the scoring still measures benchmark answers, not whether clinicians or patients benefit.

Alex: That's the boundary I would keep. The component-removal tests add supporting evidence: disabling graph exploration, image retrieval, or selective disclosure lowers accuracy on the selected diagnostic benchmarks. Those tests support complementary roles for the components, within a narrower evaluation.

Sam: Selective disclosure sounds economical. Does it actually reduce the computational bill?

Alex: It reduces retrieved evidence, not total token use. On the free-response benchmark with Gemini, SkinMind averages about fourteen hundred evidence tokens per question. LightRAG uses about seventy-nine hundred. Tokens are the units of text processed by the model.

Sam: But repeated decisions also require inputs and outputs. Does the smaller evidence bundle outweigh that extra interaction?

Alex: No: SkinMind uses roughly twice as many total tokens as LightRAG in that analysis. Earlier interactions and tool instructions contribute to repeated inputs. Its efficiency claim is about selecting compact evidence, not making the entire process cheaper.

Sam: There's another risk: a model could navigate very carefully around its favourite diagnosis. Do the failures show that happening?

Alex: They do. In one failure, exploration centres on cellular neurothekeoma, a candidate skin tumour. The model checks features supporting that candidate but never explores pilomatricoma, the correct option. In another, it favours keratoacanthoma over basal cell carcinoma while leaving growth history and central keratin unverified.

Sam: Those aren't simply missing-reference problems. The model has evidence, but doesn't complete the comparison that would make its decision defensible.

Alex: That's consistent with the authors' analysis. They suggest improving tool-use strategies to encourage more targeted, comparative exploration. Giving a model navigation tools does not ensure that it uses them to challenge its preferred answer.

Sam: Who should read the full document, and where should they start?

Alex: Researchers building medical retrieval systems or interactive graph search should read it. Start with Hierarchical Graph-Probing Interfaces to see what choices the model actually gets. Then read Main Results and Context Efficiency Analysis together, so compact evidence isn't mistaken for lower overall cost.

Sam: And the section that keeps the reliability claim in perspective?

Alex: The appendix's Failure Case Analysis. It shows the distance between collecting relevant facts and distinguishing diagnoses. For everyone else, remember this: better retrieval helps, but reliability still depends on asking what would distinguish the alternatives.

Sam: Thanks for listening.