My main research topic is the study of different human-computer interactions. In such interactions, one important task is (1) to detect and recognize the internal state of the human partner (e.g. stress level, cognitive load, fatigue, etc.) and (2) to design the user assistance that may be needed to increase the efficiency of the interaction. Such researches are carried out through the evaluation of different multimodal information sources. Examples of applied methods and smart tools are eye-tracking, saving logs of the software or the user interface, detecting head, hand and arm movements, facial expression recognition - the latter from simple off-the-shelf webcam images. Deep learning algorithms and machine vision techniques support efficient analysis.