Our company has been using Machine Learning technologies to solve various non-trivial problems since 2007. Machine Learning is a set of methods which allows to “teach” a computer how to work on certain tasks independently. We are talking about artificial intelligence, which is a big part of the new technological paradigm. For example, with the help of neural network (one of the Machine Learning techniques), one can teach a computer to recognize voice recordings or “read” handwritten texts.
The applications of artificial intelligence can be differentiated by the type of input data
- Images: It is possible to teach a computer to recognize particular people in the crowd or detect pathological changes in the tissues using X-ray or similar medical images
- Tables: Make a prediction for sport events or suggest a diagnosis to a patient based on a number of metrics
- Video Stream: Pedestrian detection and recognition of other subjects on the road such as obstacles, road signs and traffic lights
- Sound: Voice command recognition, detection of emotional aspects of the speech or recognition of a particular speaker
- Text: Intellectual search for specific information in the literature
- Logical and Scientific Problems: Playing logical games and solving math equations
- Graphs and Ontologies: Automatic translation from one language to another and detecting patterns in a large amount of events
Our company has accumulated successful experiences in solving biological problems, image processing, and voice and text recognition using Machine Learning methods. In our opinion, finding genetic and/or biometric markers as well as drug targets are the most difficult tasks to solve. Besides mathematical methods, we solve such problems using specialized knowledge databases as well as our expertise in biology and medicine. Our predictions go through a number of lab verifications and tuning cycles in order to achieve the highest level of veracity.