does not look into the paper?

[Hansyvea]

The answer given is quite shallow which gives me a feeling that it only analyses the title but not the content of the paper.

[lifan0127]

Hi @Hansyvea Thanks for your feedbac!. This will be improved with full-text based analysis scheduled for the next major release.

[Miguilo]

Hi @Hansyvea Thanks for your feedbac!. This will be improved with full-text based analysis scheduled for the next major release.

@lifan0127 thanks to the attention! I think that this major update will increase a lot of the users in the ARIA. Excited to see it in action.

[MohamedNumair]

Hi @lifan0127 , I just started using it, seems a great tool, but I also wonder if it can access the full text from PDF? because all answers seem to be based on abstract and not the pdf content.

[lifan0127]

Hi @MohamedNumair as mentioned above, this is something I am actively working on. Stay tuned 😉

[tommy-qichang]

Since the PDF file size is unclear, I think Aria might need a vector database to identify blocks useful for queries. I'm looking forward to hearing updates on this feature.