16:02:24 RRSAgent has joined #hcls 16:02:28 logging to https://www.w3.org/2024/01/18-hcls-irc 16:02:28 rrsagent, make logs public 16:02:32 Meeting: FHIR RDF 16:02:38 Chair: David Booth 16:05:15 Topic: Introductions 16:09:20 Darrell Woelk: involved in semantic world, going back to MCC. Then worked on Carno, using Cyc. The started Info Sloop, Made a company in 1995, then got into healthcare, building a soft dev platform in healthcare Started w FHIR in 2015. Green Room Tech is our latest. I do FHIR consulting, both tech and business. When Chat GTP cam out, I started working on AI again. Ivolved w U Texax Information, doing lectures on FHIR, then last hyear on gene 16:09:20 rative AI. Had 4 gra dstudents working on gen AI, and want to do more this year. 16:09:50 Topic: Darrell's ideas on student involvement. 16:10:01 darrell: Working w Dr Ding. 16:10:07 ... KG and LLM. 16:11:14 .. Spoke w Juan Sequeda in Data.world. They did soft dev and pub an article about KG and LLM bencharks. 16:11:30 ... It can gen SQL, using SQL DL. 16:12:10 ... Doing the same thing w an ont and SPARQL 16:12:22 ... Juan did a mapping into DB 16:12:35 ... They had a claims ont. 16:12:51 ... data.world executes the SPARQL queries. 16:13:06 ... They turn those SPARQL queries into SQL and return answers. 16:14:00 ... They did a benchmark, comparing SQL-only-based vs ont-informed. It was 50% better, and they showed why. 16:14:17 .. They used zero-shot prompt. 16:14:35 ... They made this available for others to try. 16:15:07 ... Proposing to do this in the healthcare domain. On the left side, that meansw we don't have to tell chatGPT what the FHIR resources are, becuase it already knows that. 16:15:19 ... Use a FHIR API earch instead of SQL. 16:16:39 .. On the RHS want to do something similar to what they did w insurance claim, but w phenotype ont. 16:17:09 ... Then using FHIR RDF data. 16:17:36 ... Or could export all my FHIR RDF data into the triplestore. 16:18:12 eric: Sounds fun. There might be some code that you could reuse, that does part of that. 16:18:56 ... SPARQL query gets dismantled to match hierarchies agains a library of FHIR path. Rule head would be FHIR Path, body would be SPARQL. 16:19:30 .. Divide up resources and test it against know FHIR paths. Launch , get results back, bind appropriate resources. 16:21:06 (Eric explains the code more) 16:21:26 eric: That's in typescript right now, Claude Nanjo and I worked on it. 16:22:33 eric: There's a shex schema for FHIR. 16:22:41 dbooth: And a FHIR ont. 16:23:32 darrell: Phenopackets include a lot of data. Vulcan project trying to come up w FHIR rep of phenopacket. 16:23:45 ... Met w them yuesterday. 16:25:18 jim: HPO folks have a transformation to FHIR ont. 16:25:33 ... They're at least workin on it. 16:25:48 ... Someone who works for Chris Chute working on it. 16:26:48 darrell: These are grad students. Class is AI and health. Want a group of 2-3 students. How to intro FHIR RDF? 16:27:42 https://yosemiteproject.org/tutorial-fhir-rdf-as-a-bridge-to-the-semantic-web-in-healthcare/ 16:28:38 dbooth: That hasn't been updated to R5 yet. 16:30:26 eric: Did the HPI queries assume that the closures implied by the ont were already instantiated in the SQL data? 16:30:30 darrell: IDK. 16:31:12 eric: could either assume no inference, or could add subsumptions in tables, and join those also. 16:31:35 ... Or you could transform the query to add the subsumptions. 16:32:14 darrell: They product is a catalog. Guessing that they're not pushing on inference, but I'm interested in it. 16:32:29 ... This would be a first step toward building a test system. 16:32:46 jim: I think the work by Joe Flack on ontology-to-FHIR has been folded into OAK python (Ontology Access Kit): https://incatools.github.io/ontology-access-kit/howtos/fhir-conversions.html 16:33:08 eric: Queries were supposed to go over their info mdel. 16:33:31 ... You could infer the structure from the DDL, or maybe from an ont. 16:33:43 ... But sounds like they used both. How did they use the ont? 16:34:20 darrell: Their product maps from KG to individual DBs. 16:36:11 eric: Could ask a FHIR query, either 1. "I'm looking for sombody w a procedure performed on the kidney", and that's turned into a query for procedure performed on a kidney, or on kidney lining, etc., using the closure over the subclasses to expand the query. 16:36:39 ... 2. Issue the original SPARQL query, but have the infra on the FHIR side, HAPI FhIR server goes to the term server to do that closure. 16:36:50 darrell: And HPO work is doing that? 16:37:45 jim: Joe Flack is the developer working on that at HPO. He works for Chris Chute at Johns Hopkins, but also working on Monarch initiative. 16:38:11 ... The OAK python toolkit does lots of things w ont. IDK if they're hosting the HPO ont someplace as FHIR. 16:42:40 eric: Wonder if you could use shex to steer GPT4. 16:43:43 gaurav: You could also give it a shex schema and tell GPT4 to "using this schema, give me ... " 16:44:24 eric: But a well-informed person still needs to read the GPT4 responses to tell it to correct them.. Maybe use shex for that. 16:45:16 eric: Will be hard to make it guess a good query. 16:45:39 darrell: Benchmarks that data.world did are not perfect, an dthey discuss them. 16:45:54 ... How to prompt it to correct the mistakes? 16:46:38 eric: What's "included" in the FHIR queries? 16:47:05 darrell: it joins other info to your results. 16:50:18 Topic: Concept IRIs 16:50:36 gaurav: The submission part of jira is down right now. Cannot edit it right now. 16:51:07 ... But I'll keep an eye on it. 16:51:33 Present: Jim Balhoff, Gaurav Vaidya, David Booth, EricP, Darrell Woelk 16:51:39 ADJOURNED 16:51:43 rrsagent, draft minutes 16:51:45 I have made the request to generate https://www.w3.org/2024/01/18-hcls-minutes.html dbooth 16:56:51 i/Working w Dr Ding/Darrell's slides: https://lists.w3.org/Archives/Public/www-archive/2024Jan/att-0001/Human_Phenotype_Ontology_and_FHIR.pdf 16:56:54 rrsagent, draft minutes 16:56:55 I have made the request to generate https://www.w3.org/2024/01/18-hcls-minutes.html dbooth