When ARTIN founders were looking for a name for the company in 1998, they came up with ARTIN – by creating an abbreviation of ARTificial INtelligence. They both did a PhD in said field of expertise and had a guess that AI would once become an important technology for commercial sector.
And the time is now, so maybe you ask what do we do in the AI area and what can we offer to our customers:
Image processing – we use AI image processing technology mainly on the Roboauto project. This is primarily about convolutional neural networks, deep learning, YOLO, and similar algorithms. We use these technologies to recognize objects from the camera image such as road lines, drivable terrain, vehicles and pedestrians detection. The image can also be used to locate the vehicle on the map and to measure odometry (measuring where the vehicle has moved and how fast it is going). We use tools such as Keras and Tensorflow.
Tesabot project also uses image recognition to analyze tested applications, progress screens, or identify individual application screen components (buttons, text boxes, links, etc.). Recurrent networks can also be used to detect anomalies and potential application errors.
Speech processing – vcelka.cz project is using recurrent neural networks. Recurrent networks identify the correct pronunciation of individual words or sentences.
Machine learning – even for smaller tasks in the Roboauto project, we use machine learning elements such as adaptive regulation with the learned model of a regulated system using neural networks.
Chatbots – to get to the thing most people might think of when you say AI – we also use natural language processing technology (NLP) in order to create communication chatbots for customer contact, as a replacement of FAQ sites and generally as a support channel with a new form of interaction using common language.