While context-awareness (as a subfield of human-computer interaction, HCI) has been defined as software that āadapts according to the location of use, the collection of nearby people, hosts, and accessible devices, as well as to changes to such things over timeā and that āa system with these capabilities can examine the computing environment and react to changes to the environmentā, we should expand this definition to reconfigurable hardware, e.g. in an IoT scenario. The pursue of this research path leads to an improved understanding of the context, including its most difficult aspects, i.e., human emotions, which has the potential to span the development of an artificial emotional intelligence and a shift from natural (intelligent) to personal (emotional) user interfaces.
Deep learning as a part of the machine learning area has been through one of the most remarkable breakthroughs, spreading its popularity to almost every research area in computer science and beyond. One of the drawbacks in the majority of modern machine learning algorithms, including deep learning, is parameter-dependent performance. There are dozens of free configuration parameters that need to be configured before fitting a model. While traditional parameter setting strategies usually do not perform well, seeking new optimisation algorithms suitable for parameter setting in advanced machine learning algorithms is an active field. Another very active field seeking new contributions is merging traditional signal processing methods with machine learning, for example in the area of biomedical signal processing .