Fomubad Borista Fondi, Azanzi Jiomekong Fidel Integrating SPARQL and LLMs for Question Answering over Scholarly Data Sources https://arxiv.org/abs/2409.18969
or query a SQL database or a knowledge graph with SPARQL. You can do it now, but it's hit-or-miss, and a database lookup is by definition very discrete and not very optimizable through gradient descent. if you had a good way to get a gradient-free optimization of these search abstractions, yikes
It's for a chap external to Parliament who's building a dashboard of energy related things. He's already consuming RSS from gov.uklegislation.gov.uk. We have a SPARQL query churning out JSON and CSV for energy related SIs, but ideally he wants RSS.
Interconnected relationships across OCLC WorldCat data can be searched in an intuitive way using the Entity Connections API. There's also a SPARQL end-point Entity Query API #eubfws2024
📣 Introducing the new BacDive knowledge graph 🦠💡, which contains more than 16.5 million triples representing so far 26 BacDive entities and can be accessed through a SPARQL ✨ endpoint. 🔗 sparql.dsmz.de/bacdive#opendata#DSMZDigitalDiversity#databases 🧪🖥️🧬
“and our SPARQL library now boasts a new query listing all SIs laid under a given Act in the last month. Not being one to leave a job half done, Jayne also added a new query listing all SIs laid by a given body under a given Act in the last month”
How good is the reasoning in the new o1preview (a/k/a “strawberry”) model from @openaidevs.bsky.social#SPARQLchatgpt.com/share/66e469...#GenAI#AI#UseCase
How good is the reasoning in the new #o1preview#SPARQLhttps://chatgpt.com/share/66e46997-cc5c-8011-bf9e-505863782431#GenAI#AI#UseCase
オルタナティブなSPARQLエンドポイントは以下のとおり。まだ全然試してないんで、今後ちゃんと勉強しないとね……。 Wikidata:SPARQL query service/Alternative endpoints www.wikidata.org/wiki/Wikidat...