Welcome!
My name is Peng Zhou. I'm a first year computer science Ph. D student at Georgia Tech in the School of Interactive Computing. I graduated in 2006 from the University of California, Berkeley, with a bachelors in Computer Science and a bachelors in English Literature. I am currently doing research in the area of Artificial Intelligence.Research Interests
Broadly speaking, I am interested in the development and application of AI and ML techniques to interactive domains such as video games, virtual worlds, and social settings. The former two are of particular interest, as traditionally virtual domains offered great opportunities for AI research because they were complicated enough to be difficult, yet not so complicated that the amount of effort required to design even a rudimentarily intelligent agent was astronomical. Yet, I am not only interested in using these domains as testbeds for algorithms and techniques in AI. I am also interested, and perhaps moreso, in the problems within these domains and how they can be solved through AI in a process of mutual advancement.
To be more specific, I am currently interested in three major problems. The first is that of creating adaptive, believable, and autonomous agents, by which I mean agents that are capable of operating in environments alongside or against humans without coming off as trite, bland, or pre-scripted. This problem, in some sense, goes to the very heart of AI, as it involves the questioning and analysis of what makes entities intelligent and/or lifelike, and how to operate such entities in environments that are potentially continuous, partially observable, and non-deterministic: a hard problem indeed.
The second problem I'm interested in is the issue of assisted authorship and content creation, which has become increasingly important as the human costs of production have grown drastically alongside quality and sophistication. Here, I am talking about such phenomena as the ramping growth in time consumption for creating art assets (ie for games and movies), the sheer complexity of constructing massive virtual systems (ie Second Life), and the need for tools that would enable people outside the academic community to construct and experiment with AI methods. I think a potentially revolutionary way of thinking about these problems is through the lens of assisted machine creativity, in which the human and the machine interact to produce original ideas and assets.
The third problem that I'm interested involves combining the first two and looking at how they can expand the design of interactive entertainment, if not altogether generate an entirely new genre of entertainment. I am inspired, in this, to look at the intersection of entertainment technology and creative design, particularly how the two exist in a state of symbiosis wherein the strengths of one could be used to offset the weaknesses of the other. One could imagine, for example, a real-time strategy game specifically designed to exploit the adaptivity offered by machine learning by challenging the player to train a squad that would then fight autonomously in different situations. The possiblities are mind-boggling, and represent a fertile ground for interdisciplinary research between science and art.
My approach to the problems draws from a variety of angles, ranging from RL, imitation learning, and planning. I plan to explore even more techniques in the future, as I think the problem, and not the method, should determine the research being done.
That said... I am personally fascinated by the idea behind autopoietic (self-creating) systems, and will likely apply them to everything I do. To me, the concept poses the best - and perhaps only - answer to the limits of human cognition in the face of exploding complexity. Alas, there are times when I think that The Answer to all of AI lies in building hardware that would efficiently and naturally support such a process, and that all our intellectual teeth grinding consists of trying to optimize fundamentally simple learning algorithms for an intrinsically flawed system.
Cookie wisdom
The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.
- George Bernard Shaw
Neither a lofty degree of intelligence nor imagination nor both together go to the making of genius. Love, love, love, that is the soul of genius.
- Wolfgang Amadeus Mozart
True knowledge is to know the extent of one's ignorance.
- Confucius