Due to the fact that the field is so new, we have not listed any specific subareas or topics. The door is open for suggested topics. We hope to form a public forum for discussions and debates!
The potential applications and economic returns on intelligent systems research have been widely recognized for many years. Several road maps and approaches have been proposed by various research groups in the open literature. Our central focus here is a proposal to design a brain-like computer. The fusion of fundamental scientific research results and state-of-the-art technology such as opto-electronics, nano- and photonics will be emphasized. One key issue facing us today is understanding how information is processed and knowledge networks are organized in the brain. Biological systems such as brains have enormous capabilities in information processing including information storage, retrieval, sensor fusion, cognition and recognition of spatio-temporal patterns. The motor controls and behaviors of biological systems are powerful and agile indeed. Encouraged by recent advancement in so called biologically inspired computational intelligence algorithms and natural computing, it is only prudent for us to push further in this direction. The consortium of useful methodologies such as artificial neural network, genetic algorithm, fuzzy logic and natural computing certainly has changed the landscape of artificial intelligence. Coupled with the amazing natural abilities of the biological systems in adaptation, learning, self-organizing, and reasoning, one should not be surprised that even more powerful algorithms that are more agile, controllable, and reliable can be created in the near future.
With all the evidence we have on hand, it is perhaps safe to say that biological systems often perform orders of magnitude better than systems based on current silicon device technology. The key to improving the performance of intelligent systems will hinge upon our ability to understand how the brain actually works! Hence it is only fair to say that the brain-like computer architecture provides a natural platform upon which to launch further research into the realization of intelligent systems. It is entirely possible that learning how the brain processes
information may lead to significant new information systems in terms of algorithms, software and systems.
Since there has been tremendous progress in opto-electronics, photonics and nano-technology, it is timely for us to discuss brain-like computer architecture in concert with advancements in hardware technology. By taking this view, the research findings may be translated into practical applications more quickly. This approach also would most likely enlist the support of the robotics sector.
One practical example of such intelligent systems is bio-inspired robotics. The research tools in this case consist of artificial intelligence, cognitive science, psychology, neurobiology, and control systems. Robotics research has shown that we can learn the function of all modalities of sensory inputs: somatosensory, visual, auditory and visceral. In fact, these ideas have been explored by many research groups around the world. Our workshop intends to provide a public forum for debate and discussions.
Another interesting example is the evolvable hardware systems, which possess self repair and self reproduce hardware. The evolutionary hardware design methodology may become a viable technique in the tool box of contemporary technology. Once again, this exemplifies the benefit society has gained from a deeper understanding of biological systems. Unfortunately, our fundamental understanding of how the brain works is still very limited! Unlike the DNA models for molecular biology, brain research has very few mathematical models to lean on. Nevertheless, we do know that the brain processes information in the form of spatio-temporal patterns. Pattern discovery has become critical in dealing with the massive amount of data. The problem of high dimensional data modeling, novel clustering techniques, combination of ensemble of classifiers, method for scaling up supervised learning algorithms, and reinforced learning all demand the ability to handle complex and deformable templates, as pointed out by NSF.
On the other hand, many other fields of research such as image processing, computer vision, speech recognition, tactile and other unusual perceptions all have extensive knowledge bases and technical know-how. They have matured enough to provide insights into the study of brain-like computer architectures. Coupled with research results in explicit, declarative, and implicit memories, or the combination of different kinds of memories, this fundamental scientific knowledge may very well be the key to developing a high performance architecture for brain-like computers.
To name only one example, we now know that its data base should be a pyramid and its memories should be distributive. As scientific research continues to bear fruit and technology advances, the public debate should grow. The incentive for computer scientists, mathematicians, engineers, natural scientists, biologists, etc. working together will also grow with it.
Finally, computational biology and genomic bioinformatics have changed the landscape of scientific research. The theory of Boolean logic, cellular automata, artificial life and the like may very well bridge the gaps between the life science and technological communities.
For the last decade, JCIS has provided a public forum in the broad field of information sciences. A shift of paradigm to data driven algorithms has been clear. It is our hope that these discussions of brain-like computer architecture will bring us to another level of accomplishment.