Doqume against COVID-19 is a free solution we are now sharing with researchers and medical professionals to help optimize their research for COVID-19 related insights from literature, news, and structured data resources.
We started constructing this research platform after joining in a collaboration with a group of Lithuanian friends and medical researchers working at the front-line of this crisis. In late March the group was tasked to help produce state-level guidelines and to engage in policymaking conversations with government institutions.
As the group kicked-off work, rapid acceleration in new coronavirus literature, new data, and news coverage has quickly emerged as key to the group’s operational efficiency.
The below aspects were particularly daunting:
- Keeping up with the pace of new research and with changes in canonical knowledge
- Keeping up and addressing the propagation of speculative or “fake news” media content
- Setting up a Knowledge management system in conditions of fast-changing knowledge
How can you use Doqume?
Today you can search for any keyword entity and our algorithm will:
1. Find publications and news matching your search;
2. Semantically analyze sentences where your searched entity was found;
3. Identify topics and entities also mentioned within that same sentence;
4. Display those entities in clusters of topics based on custom-built taxonomies to help navigate insights by categories.
Found relevant insights can be added to our smart editor to make a private knowledge base where relevant research is maintained.
Upcoming:
We will keep the below items up-to-date based on where we will be. Help us prioritize by getting in touch. Thank you!
- Alerts: we are working on a smart alerting system that can identify news based on facts rather than keywords.
E.g. be alerted when companies in a certain region start production of personal protective equipment.
- Datasets: we plan on integrating into our knowledge-graph openly available datasets for querying COVID-19 related data.
- Algorithms: In our next release we will implement algorithms that can analyze text based on underlying facts rather than identified entities.
This will help our Research Engine locate sentences where an entity, e.g. “ocular tissue”, is mentioned as a possible medium of diagnosis, versus a medium of infection.
Our idea is to assist the navigation of insights by identifying facts and finding supporting evidence on their validity.
Collaboration:
We are actively looking for collaboration in the form of feedback or partnerships.
If you want to learn more about this project, or if you have an idea on how this resource could help you, please do get in touch at vaidotas [at] doqume [dot] com.
Thank you.
Francesco and Vaidotas Co-founders of Doqume.