Dr. Isbell's lab applies statistical machine learning techniques to building autonomous agents that live and interact with large numbers of other intelligent agents, including humans.

Their technologies have applications in a wide variety of fields, including robotics, biosciences, gaming, and online business. 

The Laboratory for Interactive Artificial Intelligence (IAI) is dedicated to the idea that human-level intelligence is driven to be interactive. In other words, Dr. Isbell's work rests on the belief that intelligent agents seek out opportunities to interact and communicate with other intelligent agents and that this drive encourages learning.  

The IAI group studies how intelligent and adaptive systems interact with humans on their own terms, enabling the system to become a natural part of the human environment. The IAI group is interested in building real and robust systems that demonstrate the usefulness of this approach and in developing fundamental algorithms and technology that support such systems. 

Much of Dr. Isbell's work centers on modeling human beings and their interactions using statistical machine learning techniques to perform activity discovery, modeling, and recognition. These activities are supplemented by research on engineering and algorithmic methods for ensuring the scalability of the group's machine learning solutions.  

Research Goals 

Interactivity: Designing AI and machine learning systems that can build relationships with humans and other intelligent agents

Adaptability: Developing methodologies for building robotics that are collaborative and believable agents and that can interact seamlessly with similar agents as well as with humans

Scalability: Innovating engineering and algorithmic methods to scale up machine learning techniques

Expansion: Adapting machine learning solutions for non-expert authors, designers, and developers in a variety of fields  

Activities  

  • Activity discovery: Supporting the abilities of machine systems to gather information from their environments and make resulting decisions  
  • Modeling: Utilizing machine learning to model human behavior so that intelligent systems can better adapt to their environments 
  • Recognition: Leveraging algorithms and other technologies to design intelligent systems that can accomplish tasks more effectively and can robustly respond to changes in their environment  

Leadership  

  • Executive Director, Constellations Center for Equity in Computing, Georgia Tech
  • John P. Imlay Jr. Dean, College of Computing, Georgia Tech