Mikhail Jacob is a Ph.D. student at the Georgia Institute of Technology. He received his B.E. in Computer Science Engineering from Manipal Institute of Technology in 2011, and his M.S. in Computer Science from the Georgia Institute of Technology in 2013. He has developed AI-based technologies in areas such as pretend play, improvisational dance or movement, improvisational theater, visual art and procedurally generated video games. He is interested in researching socio-cognitive models of creativity in artistic or expressive domains and his current research is on computationally modeling embodied narrative intelligence and improvisation of embodied narratives between humans and virtual agents. His work has been featured at venues such as the Creativity & Cognition 2015 conference Art Exhibition in Glasgow, ICIDS 2014 Art Exhibition titled "Remembering/Forgetting" in Singapore, multiple Tech Arts Festival events at Georgia Tech, the DAEL Window Project in downtown Atlanta, the Autumn School on Computational Creativity in Porvoo (Finland), the Georgia Game Developers Association in Atlanta, and at the AIIDE 2013 conference in Boston.
Computational Play Project
The Computational Play Project seeks to model playful behavior computationally in software agents and in robots, in domains ranging from object-based pretend play to playful human - agent collaboration on tasks.
Computational Play Project
Within the project, my work has focused on understanding the act of pretense as a layering of a pretend world (diegetic) frame onto a real world frame. This layering is accomplished through a conceptual blending process we named object blending. Through object blending a bottle being raised above the table can be recognized as a space shuttle launching, and conversely a vertically standing bottle can be chosen as a potential toy to substitute for a space shuttle about to launch in the pretend world.
The Viewpoints AI installation (at the Adaptive Digital Media Lab) represents an attempt to create a fully-realized interactive narrative experience that is absent of any pre-defined instantial content by the authors for example pre-authored story points. As opposed to focusing on author-created content that reflects their specific view of an experience, we have instead created a movement-based play space where interactors can freely dance and move with a virtual AI-based character named VAI. VAI remembers and uses all its interactions to help inform future narrative decisions. While the Viewpoints AI installation has little concept of an explicit “story”, narrative content (which refers here to coherently ordered events) is defined by procedural representations of the Viewpoints movement theory (from theater and dance) as well as memories of its past interactions with people. VAI uses this procedurally and experientially realized narrative content to present an engaging movement-based installation that any audience member can walk up to and immediately begin playing with, while contributing to the ongoing narrative tapestry of movement that VAI produces with people.
Improvisation is creativity at its most present and impressive. Whether in jazz performance, or improvisational theatre, or improvising a new plan of action when something goes awry. The Digital Improv Project (at the Adaptive Digital Media Lab) sought to understand the cognitive processes behind improvisation by studying improvisational actors from around Atlanta. These studies were then used to build models of improvisation within a limited Wild West domain so that improvisational actors could approach the system and play improv games with the virtual performer on screen such as Party Quirks and Three Line Scene. The system was even accepted as a performing troupe to the Chicago Improv Festival.
The Game Adaptive Intelligent Agents (GAIA) Project (at the Design Intelligence Lab) is an agent design system that creates agents containing a model of their own behavior. It combines CAD for agent design with self-adaptive agent intelligence to enable rapid prototyping of agent designs. GAIA determined adaptations of its behavior model for future success after a past failure as well as for successful execution in modified scenarios with new rules and game mechanics.
