Social virtual actors need to interact with users emotionally, convincing them in their ability to understand human minds. For this to happen, an artificial emotional intelligence is needed, capable of believable behavior in real-life situations. Summarizing recent work of the authors, the present paper extends the general state-of-the-art framework of emotional AGI, using the emotional Biologically Inspired Cognitive Architecture (eBICA) as a basis. In addition to appraisals, other kinds of fluents are added to the model: somatic markers, feelings, emotional biases, moods, etc. Their integration is achieved on the basis of semantic maps and moral schemas. It is anticipated that this new level of artificial general socially emotional intelligence will complement the next-generation AGI, helping it to merge into the human society on equal with its human members.

This work is devoted to the development of a concept-proof prototype of a special kind of a virtual actor - an intelligent assistant and a partner, called here “Virtual Listener” (VL). The main role of VL is to establish and maintain a socially-emotional contact with the participant, thereby providing a feedback to the human performance, using minimal resources, such as body language and mimics. This sort of a personal assistant is intended for a broad spectrum of application paradigms, from assistance in preparation of lectures to creation of art and design, insight problem solving, and more, and is virtually extendable to assistance in any professional job performance. The key new element is the interface based on facial expressions. The concept is implemented and tested in limited prototypes. Implications for future human-level artificial intelligence are discussed.

A weak semantic map, as opposed to a strong semantic map, allows for a choice of coordinates that are characterized by definite semantics: e.g., valence, arousal, dominance. Weak semantic maps of words can be built from synonym-antonym dictionaries, by pulling synonyms together and antonyms apart. Polysemy is one of the problems with this approach. Indeed, typically one and the same word has multiple meanings, while it has to be represented by only one point on the map. To solve this problem, it seems natural to use word senses rather than words as the map elements. In this work, we consider a semantic map of word senses built from the thesaurus of Microsoft Word for the Russian language. To determine senses of words with the purpose of its further representation on the built map of word senses is developing a method, the result of which is presented in this article. The main conclusion is that semantic maps of word senses have an advantage over semantic maps of words, since they allow for a context-specific evaluation of word meaning. This approach removes the restriction caused by polysemy - the multiplicity of senses of one word.

Cognitive psychology has accumulated a vast amount of knowledge about human social emotions, emotional appraisals and their usage in decision making. Can an emotional cognitive architecture injected into an artifact make it more "humane", and therefore, more productive in a variety of creative collaboration paradigms? Here, we argue that the answer is positive. A large number of research projects in the field of digital art that are currently underway could benefit from integration of an emotional architecture component into them. An example is the project Robodanza (a robotic dancer), the functioning of which is based on a hidden Markov model trained by a genetic algorithm, yet lacking deep emotional intelligence. Generalizing on this example, we outline a roadmap to building a variety of useful virtual creative assistants to humans based on an emotionally intelligent cognitive architecture.