Our team at https://github.com/Cinnamon/kotaemon/ has been working on a public demo to showcase the new advanced citation features in our RAG (retrieval-augmented generation) application.
We’re excited to share a web app that lets users explore top daily machine learning (ML) papers on Arxiv (via the HuggingFace API) and upload their own Arxiv papers to get LLM-assisted summaries, mind maps, and answers to questions based on the content.
Some notable features:
- Instant Summaries & Mind Maps: Generate concise summaries and visual mind maps for any Arxiv paper.
- Transparent Citations: Verify AI-generated answers with clear, evidence-backed citations. Citations are highlighted directly in the in-browser PDF viewer.
- Flexible Citation Options: Choose between highlights and inline citations. Plus, click on any sentence in the AI-generated response to see its supporting source from the original paper.
- Multi-Paper Analysis: Compare, contrast, and compose summaries from multiple papers simultaneously.
- Complex Question Solving: Use Chain-of-Thought (CoT) reasoning mode to break down and solve complex questions step-by-step.
- Customizable & Private Hosting: Easily self-host or customize your private app via HuggingFace Spaces. You can securely connect your LLM and upload your own document collections.
We’d love to hear your thoughts, feedback, and recommendations as we continue improving this tool.
Hi HN,
Our team at https://github.com/Cinnamon/kotaemon/ has been working on a public demo to showcase the new advanced citation features in our RAG (retrieval-augmented generation) application.
We’re excited to share a web app that lets users explore top daily machine learning (ML) papers on Arxiv (via the HuggingFace API) and upload their own Arxiv papers to get LLM-assisted summaries, mind maps, and answers to questions based on the content.
Some notable features:
- Instant Summaries & Mind Maps: Generate concise summaries and visual mind maps for any Arxiv paper.
- Transparent Citations: Verify AI-generated answers with clear, evidence-backed citations. Citations are highlighted directly in the in-browser PDF viewer.
- Flexible Citation Options: Choose between highlights and inline citations. Plus, click on any sentence in the AI-generated response to see its supporting source from the original paper.
- Multi-Paper Analysis: Compare, contrast, and compose summaries from multiple papers simultaneously.
- Complex Question Solving: Use Chain-of-Thought (CoT) reasoning mode to break down and solve complex questions step-by-step.
- Customizable & Private Hosting: Easily self-host or customize your private app via HuggingFace Spaces. You can securely connect your LLM and upload your own document collections.
We’d love to hear your thoughts, feedback, and recommendations as we continue improving this tool.
Check out the demo here and happy hacking!