For action recognition learning, 2D CNN-based methods are efficient but may yield redundant features due to applying the same convolution kernel each frame. Recent efforts attempt capture motion information by establishing inter-frame connections while still suffering limited temporal receptive field or high latency. Moreover, feature enhancement is often only performed channel space dimension ...

Computer-based learning environments can produce a wealth of information on each student action, which can often be represented at multiple levels of abstraction and with a variety of features. This paper extends an exploratory sequence mining methodology for assessing and comparing students’ learning behaviors by autonomously identifying abstraction levels in a hierarchical taxonomy of actions...