How a Consensus Search Works
This article will lift up the hood on how we go about delivering you results in our beta product. Remember, our product will always be a work in progress, and this is just the first iteration.
The current Consensus product is a search engine that uses AI to instantly extract and aggregate scientific findings related to a user’s query.
At a high level, our product follows these steps to produce results:
The user enters a query.
Our AI searches through our database of hundreds of millions of scientific papers for a subset of relevant papers based on a standard keyword search.
From these relevant papers, the Consensus model “reads” the papers for you, extracting sentences when authors state their findings based on evidence.
*To accomplish this, we trained our AI models on tens of thousands of papers that have been annotated by PhD’s.
We then deliver a list of these findings sorted by an algorithm that combines how likely it is that the result addresses your query (relevance), with the quality of the result (how it scored on our extraction model), and then small weights to citation count, and publish date.
The results that are returned are all findings pulled directly from scientific research.
Because we are not generating any text ourselves, it is unlikely we are outright misrepresenting what a scientist is putting forward, however, context is always important, and we encourage everyone to click on the results that interest them.
This page was created based on feedback from our first users. If there is more you want to learn, please email us with your feedback!
Additionally, if you want to see other content we have created to help improve our user experience, please check out our piece on Consensus Best Practices.