DDMAL/linkedmusic-queries

For a one-shot SPARQL generation prompt trial, we need to have a storage of pairs of NLQ&corresponding correct SPARQL

priority: low

#54

· candlecao opened

on Feb 12, 2025

CTM: when you prompt GPT to generate SPARQL, if the natural language question is expressed in Chinese, it usually defaults to add <code>@zh</code> sign and related code, then the desired result cannot be retrieved

priority: low

#48

· candlecao opened

on Jan 25, 2025

CTM: for 朝鲜族民歌、潮尔（Tsuur）在我馆是否有对应的特藏资源收录？(translated as: Does our library have any special collection resources on Korean folk songs and Tsuur?)--the entity in the question is ambiguous that it sometimes cannot generate the expected SPARQL

priority: low

#46

· candlecao opened

on Jan 24, 2025

Without "inverse properties" in the ontology vocabularies, ChatGPT struggles to effectively identify properties in more diverse NLQs (CTM)

medium

#42

· candlecao opened

on Jan 22, 2025

For NLQ "四川扬琴是否有代表性的曲目，这些曲目可能用到了哪些乐器？"(translated as "Does 'Sichuan Yangqin' have representative pieces, and what instruments might be used in these pieces?")--chatGPT cannot generate a perfect SPARQL

medium

#41

· candlecao opened

on Jan 18, 2025

ChatGPT cannot clearly differentiate between class and instance in the given natural language question

hold on

medium

#34

· candlecao opened

on Jan 9, 2025

We are unable to rehearse the VoID-generater toolkit provided from the paper "LLM-based SPARQL Query Generation from Natural Language over Federated Knowledge Graphs"

medium

#32

· candlecao opened

on Dec 15, 2024

We lack an overall context (e.g., ontology) for MusicBrain database.

priority: low

#26

· candlecao opened

on Nov 22, 2024

Not able to query lots of strings in RDF(TheSession)

medium

#9

· candlecao opened

on Sep 8, 2024

Look into the search function on the webpage of MusicBrainz

priority: low

#4

· candlecao opened

on Aug 7, 2024

Look into why we cannot upload the RDF of "The Session" database.

medium

#2

· candlecao opened

on Aug 5, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

For a one-shot SPARQL generation prompt trial, we need to have a storage of pairs of NLQ&corresponding correct SPARQL

CTM: when you prompt GPT to generate SPARQL, if the natural language question is expressed in Chinese, it usually defaults to add <code>@zh</code> sign and related code, then the desired result cannot be retrieved

CTM: for 朝鲜族民歌、潮尔（Tsuur）在我馆是否有对应的特藏资源收录？(translated as: Does our library have any special collection resources on Korean folk songs and Tsuur?)--the entity in the question is ambiguous that it sometimes cannot generate the expected SPARQL

Without "inverse properties" in the ontology vocabularies, ChatGPT struggles to effectively identify properties in more diverse NLQs (CTM)

For NLQ "四川扬琴是否有代表性的曲目，这些曲目可能用到了哪些乐器？"(translated as "Does 'Sichuan Yangqin' have representative pieces, and what instruments might be used in these pieces?")--chatGPT cannot generate a perfect SPARQL

ChatGPT cannot clearly differentiate between class and instance in the given natural language question

We are unable to rehearse the VoID-generater toolkit provided from the paper "LLM-based SPARQL Query Generation from Natural Language over Federated Knowledge Graphs"

We lack an overall context (e.g., ontology) for MusicBrain database.

Not able to query lots of strings in RDF(TheSession)

Look into the search function on the webpage of MusicBrainz

Look into why we cannot upload the RDF of "The Session" database.

Issues

Search results