Artificial
Intelligence: We concentrate primarily on learning which is acquired in
real time and grows cumulatively--Cumulative Learning(or Lifelong Learning).
We prefer biologically inspired ideas as our research basis. A current goal is
to study the hypothesis that the accumulation of experience results in
improved survival probability.
SELF-AIMING CAMERA : This is a robotic head which is driven by a neurologically
accurate model of the superior colliculus originated by my colleague, Thomas
Anastasio. The SAC head itself currently uses audio (4 microphones) and
video(one color camera) to sense the environment. The superior colliculus
model decides the "preferred target" in the scene to point to at each moment.
The 'Preferred Target' calculation is rather complex, partially learned,
depending on both simple audio and video as well as a Utility (usefulness)
function. Utility functions are our expression of the special attentional bias
of the Agent(embodied, as it were, in the model) which is time-dependent.
The model is naturally hospitable to a wide range of sensor modes, paralleling the
wide range of sensory systems found in the natural animals (e.g., somatosensory,
infrared, sonar).
Hu, C., Ray, S. R., Zhang, N.," A Model of Spatial Frequency Column in
Primary Vision System", Proceedings of SPIE Symposium on Aerospace Sensing and
Dual Use Photonics, v.2488 (Orlando, FL, April 1995).
Kargupta, Hillol and S. R. Ray, "A Temporal Sequence Processor Based on
the Biological Reaction-Diffusion Process", Proc. of IEEE Int'l Conf. on
Neural Networks, June, 1994.(Orlando).
S. R. Ray, "Integrating Knowledge from Physical Signals", Int'l. Conf. on
Integration of Knowledge and Neural Heuristics, May 9-10, 1994.(Pensacola
Beach,FL).
W. D. Lee and S. R. Ray, "Probabilistic Inference", Proceedings of 3rd
Annual Conference on AI Appli., pp. 16-20, Orlando, FL, (Feb. 1987).
S. R. Ray, W. D. Lee, C. D. Morgan, W. Airth-Kindree, "Computer Sleep
Stage Scoring--An Expert System Approach", International Journal of Biomedical
Comp., 19, (1986) 43-61.
W. D. Lee and S. R. Ray, "Probabilistic Rule Generator: A New Methodology
of Variable-Valued Logic Synthesis", Proceedings of 1986 IEEE International
Symposium on Multiple-Valued Logic (May, 1986).
W. D. Lee and S. R. Ray, "Rule Refinement Using the Probabilistic Rule
Generator", Proceedings of the 5th National Conference on AI (AAAI-86), volume
1, pp. 442-447, August, 1986.
