Gerard Assayag: Symbolic Interaction with Creative Agents

Gerard AssayagIRCAM, Paris
Improvisation can be seen as a major driving force in human interactions, strategic in every aspect of communication and action. In its highest form, improvisation is a mixture of structured, planned, directed action, and of hardly predictable local decisions and deviations optimizing adaption to the context, expressing in a unique way the creative self, and stimulating the coordination and cooperation between agents. An invaluable observation deck for understanding, modeling and promoting co-creativity in a context of distributed intelligence, Improvisation is an indispensable ability that any cyber-human system should indeed cope with in an expert way. Improvisation is instantiated in its most refined form in music, where the strongest constraints govern the elaboration of highly complex multi-dimensional, multi-scale, multi-agent actions in a cooperative and timely fashion so as to achieve creative social and cultural cooperation.
Setting up powerful and realistic human-machine environments for improvisation necessitates to go beyond the mere software engineering of creative agents with audio-signal listening and generating capabilities, such as what has been mostly done until now. Gerard Assayag has coined the term symbolic interaction in order to account for a richer and more versatile musical dialog between machines and humans, spanning over several levels of information from acoustic signals to the higher symbolic integration scales that are familiar to trained musicians. In order to foster realistic and artistically interesting behaviours in creative software agents, and to communicate with these in a productive way, we bring into synergy a combination of technical means: machine listening — extracting high level features from the musician signals and turning them into significant symbolic units ; machine cognition — discovering and assimilating on the fly structural schemes and learning powerful representations ; style modelling— elaborating consistent model of style an activating generative processes ; symbolic music representation — linking to formalized representations underpinning organized musical thinking s;a. analysis and composition. All these means cooperate — or compete — in order to elaborate multi-dimensional and multi-scale memory models, underlying processes of discovery, understanding, adaptation, reactivity, decision, planning, generation and contributing to the emergence of creative musical agents acting individually or in social interaction with other agents.
Improvised creative interactions between humans and machines can thus be studied from different perspectives (social sciences, AI, signal processing, control theory, robotics) and modalities in order to highlight the conditions under which these interactions can emerge, their temporal adaptation dynamics, their logical strategies, and to exploit their creative potential. We have proposed the expression co-creativity between human and artificial agents in order to emphasize the fact that creativity in this case is an emerging phenomenon resulting from complex interactions and complex crossed feed-backs between actors, and cannot be reduced to one or the other agent’s production in isolation, whether the latter be human or artificial.
This neutralizes the endless philosophical question of whether artificial entities can be qualified as “creative” or not, and shifts the research interest to the exploration of how to build the best possible technical tools in order to allow co-creative interaction to occur.
Symbolic Interaction with Creative Agents
Concert Monday, April 22nd at 7 p.m.
Gerard Assayag with Guest Musicians
The Loft
Free

Talk, Tuesday, April 23rd at 8 p.m.
Guest Speaker Gerard Assayag
Conrad Prebys Music Center Recital Hall
Free