The exponential growth of the power of computers in recent decades has made it possible to automate more and more tasks previously reserved for people. With machine learning and artificial intelligence this computing power can now also be used for more ‘human’ tasks. Self-learning algorithms are used, for example, to assess medical photographs, to understand speech, and to detect fraud.

Dialogic has an extensive collection of machine learning and AI algorithms to tackle complicated issues. We are therefore able to use state-of-the-art techniques in our research and consultancy projects.

Computer vision

Dialogic uses AI to classify images. The technology can, for example, be used to automatically divide photos of damaged street furniture that citizens send to the municipality and deliver them to the right person. We use state-of-the-art algorithms. On the basis of images it is also possible to recognize objects. The example below uses a model with which objects can be divided into a thousand categories. Please enter your own image or use your webcam!

Understanding natural language

Dialogic uses AI to obtain structured information from large quantities of text. A simple example is the classification of education vacancies by category. The titles of vacancies do not follow a fixed structure and often contain noise. With a simple neural network, we can classify the vacancies based on title by categories, with more than 89% accuracy. Try it below yourself!

Text mining

In order to automatically understand text, an algorithm needs to contain a proper understanding of language. The example below shows a model that we use, and that is able to translate relationships between Dutch words to useful constructs.