15-411/911 Special Topics in HHI           Interfacing with Carbon-Based           Autonomous Agents in Real-Time
		Abstract: The complexity and unpredictability of carbon-based agents has made HHI one of the most difficult concepts for computer scientists to grasp. This course will introduce students to the fundamentals of this often-overlooked field, including the pioneering research being conducted here at Carnegie Mellon University. Students will learn the algorithms and heuristics that form the basis of multi-agent interaction in a real-time environment.

Topics Covered:

• A thorough study of past research in the field, focusing on learning from the shortcomings of previously implemented multi-agent architectures, including (but not limited to):
            - Decision Oriented Robotic Communication Systems (DORCS)
            - Goal-directed Environmental Evaluation Qualifiers (GEEQs)
            - Neural Net-based Evolving Reactive Decision-trees (NNERDs)
  
  
• Various approaches to agent-level communication and cooperation.
  
  
• Examination of existing commercial off the shelf tools that can be used to enhance agent interactions, including:
            - Systematic Olfactory Analysis Package (SOAP)
            - THREADS - a fully integrated, modularly reconfigurable, extensible interface design tool
  

Lab Exercises:

• Real-time context switching
• Scalable architectures for dynamically changing multi-agent domains
• Protocol incompatibility detection and fault-recovery
• Developing context-specific sensitivities to satire

Prerequisites:

• Familiarity with natural-language parsing
• No experience with dynamic agents or agent interaction necessary

Related Courses in HHII

• 15-469 : One-to-One Interaction Between Heterogeneous Agents
• 16-43D : Coordinating Navigation in 3-Dimensional Spaces
• 15-101 : Theoretical Foundations of HHI - Complexity, Computation, and Inebriation