search

wyl-willing / MindMap

MindMap: Knowledge Graph Prompting Sparks Graph of Thoughts in Large Language Models
182 stars 19 forks source link

Path-based Evidence Graph Exploration & Neighbor-based Evidence Graph Exploration #5

Open jw9603 opened 4 months ago

jw9603 commented 4 months ago

Hello.

As the title suggests, could you please let me know the location of the code for path-based subgraph generation method and neighbor-based subgraph generation method?

Thank You!!

wyl-willing commented 4 months ago

Sure, you can find the code starting at line 505 and line 611 in MindMap.py. After that, you can get path-based and neighbor-based chains, and then you can call GPT API to transfer the chains into natural language. If you still have any questions, please let me know.

jw9603 commented 3 months ago

I have one more question: When comparing the embeddings of Mset and KG entities when linking entities, how did you obtain the embedding values about Mset and KG entities?

wyl-willing commented 3 months ago

First, you should obtain the embedding files by pre-training. For example, you can see encode_keyword_entity.py and word2vec.py in 'pre-training/chatdoctor5k'. Then use it in the MindMap.py. The details are shown in the line 481-583. [image: image.png] If you still have any questions, please let me know.

Jiwon Jeong @.***> 于2024年3月13日周三 21:49写道：

I have one more question: When comparing the embeddings of Mset and KG entities when linking entities, how did you obtain the embedding values about Mset and KG entities?

— Reply to this email directly, view it on GitHub https://github.com/wyl-willing/MindMap/issues/5#issuecomment-1994451613, or unsubscribe https://github.com/notifications/unsubscribe-auth/APCDU6ZPVFF3OSDYID4JYBDYYBKPHAVCNFSM6AAAAABDYPAZSWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOJUGQ2TCNRRGM . You are receiving this because you commented.Message ID: @.***>