Outmind Labs


Our solution is able to search and classify information intelligently.
We are working to improve the state of the art on several R&D topics,
especially on semantic research topics.
Semantic Search
We develop proprietary models to address the issues we face. Thanks to semantic search algorithms, it is possible to process the query in a vectorial way, in order to propose advanced resultsw. These models must be multilingual, and meet important scalability issues.
Detection of interlocutors
By making a query on a topic, the user has an automatic feedback of information with a high level of granularity; but Outmind goes further, in order to take into account all the users' contributions, so as to identify the best contributors on a searched topic (project, customer...).
Question Answering
In addition to providing information on documents (author, date, title...) and highlighting the best contributors according to the query, we wish to make it possible to give a precise answer to a question in natural language. The system must therefore be able to understand the query but also the corresponding response element, within elements of various types; mail, documents, pages...
Automatic labelling
Outmind develops an intelligent rules system, based on artificial intelligence, which allows to label and categorize elements automatically. The objective is to classify large volumes of data, without requiring manual action.