top of page

Our Technologies

Dynaccurate AI is the product of eight years of research and investment. In this time, we have also added additional technologies to increase utility and efficiency for all users. Check them out below!

Doctor and Patient graphic
Load Balancing.png

DyMap is a new no-code solution that any end-user can use to easily map database values for system integration. You just need to select the two database tables, click on ‘start working’, verify mappings, make corrections and push to master. You can also auto-ingest by choosing the relevant headers of a database or .csv to be uploaded. Convinced? Just use the buttons to get started!

Dynaccurate is our flagship AI. A highly sophisticated expert systems AI, it was developed over the course of eight years to manage terminologies and vocabulary alignment and harmonization.  It is a generic engine that can ingest mapped terminologies and vocabularies and can be used with multiple languages. It is particularly strong with clinical terminologies such as SNOMED and ICD, but also works well with industry vocabularies .

DyPharm is an advanced search technology for mapping drug concepts across different sources, for example across hospital formularies. It was developed to match drug identifiers to single authoritative references, primarily UK hospital formularies to dm+d. The same technology can be used to match concepts, for example between vocabularies or ontologies, to support the use of Dynaccurate AI.

Dynaccurate GUI is our web-based interface, which will incorporate all of the above technologies in one single resource location for end-users. Currently we are finalising the DyMatch, Dynaccurate AI and Mapping editor into single workstream, with a logical progression from concept matching, mapping and remapping, to allow complete semantic interoperability for hospitals and care sites

DyLink is a Named Entity Recognizer and Entity Linking solution that employs Machine Learning to detect and disambiguate mentions of concepts within free text. It can be geared towards any topic, provided that it can learn from a previously annotated corpus with texts from the domain it will operate. 

bottom of page