January 6, 2021
Tali is a Natural Language Processing (NLP) engine that is powered by medical references and guidelines (unstructured data) and acts as a question answering interface from that knowledge base. Today we are applying this engine for physicians and patients as well as medical device companies and consultants among others.
In this post, I want to tell you a little bit of a back story. How we came up with the idea of Tali for medical information and our journey to date.
In 2018, Public Health Agency of Canada (PHAC) posted a challenge and a call for proposal that asked for a solution that would help Healthcare Providers (HCP)s navigate their resources and specifically the Canadian Immunization Guide (CIG) faster.
Canadian Immunization Guide is a live five chapter book about everything immunization and vaccination in Canada. Healthcare providers are encouraged to check this reference for the most reliable and updated information at the point of care. But navigating through so much content takes so long that most HCPs prefer to Google their questions or check American resources for a quicker answer.
Evenset was awarded to conduct research and feasibility studies on state of the art NLP algorithms to find a solution for this challenge. Over the course of this project, we experimented with all major information retrieval algorithms including document retrieval, section retrieval, reading comprehension, BERT, ELMO, and many more for a question answering interface that can receive users’ questions in natural language and find the shortest answer directly from the corpus.
We built a prototype of the solution against CIG for PHAC and the results were promising, therefore we continued our research on other use cases of our engine for the healthcare and medical industry. We knew most of our knowledge in medicine is in the form of free-text, from books and references to doctors notes and government guidelines. A quick and reliable access to evidence-based information is hence a must.
Our research opened a lot of opportunities for intelligence search in pharmaceutical companies and medical device companies in addition to patient education and engagement solutions. Whenever there is a load of unstructured information and a quick access to right documents is needed, our expert NLP engine can come into play.
We registered Tali as a spin off of Evenset back in January 2020 and our story as an independent product and company started then! Feel free to check out different use cases of our engine in the healthcare and medical industry and reach out to us if Tali can make your life easier too!