Science Letter: An AI personalized Science Newsletter

Proposals Round 10
#1

Proposal in one sentence:

Science Letter is a personalized Newsletter to easily keep up to date with research on any topic, in any language and in any format by leveraging AI to adapt content based on reader preference.

Description of the project and what problem it is solving:

Keeping up to date with research
Nowadays the enormous quantity of research papers that are being published everyday makes it really hard to keep up to date. It takes a lot of time to find and curate the research content that fits ones own interests. With Science Letter, one can quickly mix and curate the list papers of interest, even for very specific topics, and quickly grasp the content of the paper in the summary format of preference. Moreover, all preferences can be configured in natural language.

Increasing the reach of scientific insights
Two problems with research papers are that they are usually written in English and for a technical audience. By using AI, we can generate content in most of the world languages and refactor the content for audiences of any level of technical knowledge.

Note that Science Letter leverages AI for web scraping, semantic search, translation and customized content generation, automating the whole crafting workflow.

This is just a starting point, would love to hear ideas and suggestions to iterate it!

Grant Deliverables:

  • An improved version of Science Letter including more Journals sources (v0 only includes Medrxiv).
  • An improved Semantic Search engine on top of a VectorDB.

Squad

Additional notes for proposals

  • Check the video of v0 demo here
  • Featured on Ben’s Bites AI Newsletter (Feb 09)
  • We have a bunch of subscribers already, subscribe here!
2 Likes
#2

I think this is a really interesting project and allows a lot of personalization, it will save time. I think the same functionality could fit in other use cases. The demo looks really promising, waiting to know more about it.

1 Like
#3

Just checked out the v0 demo , it does unlocks many possibilities and growth,
just when the translation happens, it would be a necessary thing to get the proper translation sentence making complete meaning and logic, hope you achieve that !

1 Like
#4

Thank you @localaization! Indeed, I’m trying to build it as many general as possible, so as you suggest some parts could be reused for other use cases. I will keep you posted!

1 Like
#5

Glad that you liked it @divyanshh !
As you say, the fact that LLMs are so flexible, make this approach very scalable and generalizable.
On the part of translation, I was thinking of adding an extra step to review/improve the content so it’s more coherent and logic. Thanks for the suggestions :slight_smile: