zjukg / kg-llm-papers Goto Github PK

Please add the Arxiv link to the "Knowledge Graphs Meet Multi-Modal Learning: A Comprehensive Survey" publication under the surveys section . Here is the link "https://arxiv.org/abs/2402.05391".

I just want to ask all contributors here about how you gather these LLM-KG research papers, since many papers is very fresh and up-to-date. I think it would be great to have a discussion about it (i did not see a discussion tab here, so, i made an issue in this repo instead). Here is what i want to ask:

1. The sources or methods you use to find up-to-date research papers. This could include online databases, websites, newsletters, academic journals, or any other resources you find useful.

2. If you use any specific tools, scripts, or search queries to streamline the process of discovering new papers, please provide details or examples.

3. Any tips or best practices you have found effective in finding the most relevant and up-to-date papers.

with that, i think it would help all of us to take a look and update new papers related faster.

Thank you all in advance.

Hi there,
We just proposed MarkQA: a large scale KBQA dataset with numerial reasoning, which is accepted by EMNLP 2023.
We also provided a e step-by-step symbolic reasoning path called PyQL in the MarkQA to help facilitate the multi-hop and numerical reasoning.
Here is our paper: https://arxiv.org/abs/2310.15517
Thank you for your time!

Hello！Your work is so great! We have an article about the combination of knowledge graphs and Foundation Models: R3-NL2GQL: A Hybrid Models Approach for Accuracy Enhancement and Hallucinations Mitigation. It is a method that uses a combination of larger and smaller models to generate graph database query language and retrieve the graph context required for the LLM, which is similar to your research direction.
The address of the paper is: https://arxiv.org/abs/2311.01862
The open source address for the code and dataset is: https://github.com/zhiqix/nl2gql
If you are interested, welcome to provide valuable feedback! Thank you for your time!