We propose to investigate the connection between emotions and cognition in intelligent systems through the dynamic concept of language, which links context to logic in both human and machine language. For this, our approach is inspired on aspects of the information theory of Abraham Moles. We analyze emotions under the semantic dimension, linked to a subjective context, which gives rise or not to decisions. We demonstrate that intelligent systems can, on the one hand, work with previously categorized emotions (say in a frozen context); or, on the other hand, process information under a dynamic aspect. This is possible when considering that the algorithm, as the core of the system’s language, must be adapted to functions that reflect an updated context. Thus, adapting emotions to AI means working with time-dependent communication-interpretation, in an optimized way, uniting syntax and semantics in the intended behavior of the machine. We conclude that misinterpretations can be avoided by inserting a contextual appreciation together with a categorized appreciation of emotions at the heart of the system. This allows it to absorb pre-established values in a unified way with the fluid values of emotions, making the system more intuitive. It is believed that, in this way, Computational Linguistics is focused on the characteristics of Cognitive Computing, teaching the system to interpret the appropriate context of the emotion at stake.
Motta Monte-Serrat, Dionéia and Cattani, Carlo
"Applicability of Emotion to Intelligent Systems,"
Information Sciences Letters: Vol. 11
, PP -.
Available at: https://digitalcommons.aaru.edu.jo/isl/vol11/iss4/14