UNIVERSITY OF HERTFORDSHIRE
             COMPUTER SCIENCE RESEARCH COLLOQUIUM

                            presents

   "Modeling Human Memory for Intelligent Agents and Robots"


                      Dr. Wan Ching Ho
  (Adaptive Systems Research Group, University of Hertfordshire, UK)


                    21 March 2012  (Wednesday)
                            1 -2  pm


                  Hatfield, College Lane Campus
                      * * Room C402  * *

                  Everyone is Welcome to Attend

                  Refreshments will be available


Abstract: Since decades ago the idea of creating computational
representation of experience for agents has been mentioned often in
cognitive modelling literatures. On one hand, researchers in the early
90's argued, in the context of the Turing Test, that it is questionable
whether any computer in the future can pass the Test without the ability
to experience. On the other, embodied AI emphasizes the on-going
interaction between agents and their environment, in which object
representation evolves from the experience of the agents with these
objects. Here the term "experience" is linked to the whole cognitive
concept of human "organic" memory, e.g. events attributed with "emotion"
and "meaning". It includes a range of cognitive processes that human
memory operates effortlessly: perceiving, encoding, storing, retrieving,
generalizing and forgetting of events.

In this presentation I will first talk about the current research trend
of modelling features of human memory to create useful applications
in AI and robotics domains, showing how researchers from AI and
Cognitive Science utilize characteristics of essential components
including short-term/working memory, semantic knowledge and episodic
experiences. Following that I will introduce a few research projects where 
I have been personally involved in the development of computational memory
models for believable virtual agents and social robotic companions. These
models enable virtual characters and robots to remember events that are
relevant or significant to the agent itself or to the user. This remembering 
process, or so called memory retrieval, is crucial for them to adapt 
themselves in the environment as well as in the interaction cycles with 
human users. Finally, in each of the project users can also benefit from 
the visualization of episodic events, which are chronologically sequenced
and derived from agent/robot's interaction histories.


---------------------------------------------------
Hertfordshire Computer Science Research Colloquium
http://cs-colloq.stca.herts.ac.uk