Several decades ago, at the dawn of the appearance of programs and websites, there was a need to develop a convenient user interface. Since then, templates and concepts verified over the years have appeared that make the interface of various programs convenient for most users. However, the rapid development of machine learning in recent years has prompted the idea of implementing a machine interface in such a way that it adapts to a specific user or a whole group of users.
The talk is inspired by Rahnama H., Alirezaie M., Pentland A. A Neural-Symbolic Approach for User Mental Modeling: A Step Towards Building Exchangeable Identities // AAAI Spring Symposium: Combining Machine Learning with Knowledge Engineering. – 2021.
In the first part of the report, the concept of neurosymbolic integration will be considered. This concept was proposed to combine symbolic (logical) and sub-symbolic (neural network) data processing methods.
Then we will talk about the neurosymbolic modeling of the user's consciousness. In this part of the report, we will touch on the possibility of using decentralized data, a digital model of the ontology of consciousness (if it is possible at all), and a way to update the knowledge model.
In the end, we touch on the debatable issue of the applicability of the concept described above.
Speaker: Filatov Anton Yurievich, assistant of the department MOEM SPbGETU "LETI"
Registation via link