Speaker
Description
With the overwhelming impetus of “machine learning” (some would say “hype”), HPC has exposed new opportunities for data analytics both supervised and unsupervised. But, Neural-Net algorithms are not exactly new as they were developed in the 1980s. Nonetheless, with the advent of contemporary GPU-based supercomputers large-scale informatics has emerged as a practical tool for input-stream filtering, interpolation, and clustering of massive data sets; all useful for data analytics. Unfortunately, the inchoate aspirations of the six-decades pursuit of “Artificial Intelligence” is either not understood in today’s context or is largely ignored in the stampede towards short-term papers and markets. The gratification of near-term success distracts from the far more important and essential long-term accomplishments of real Machine Intelligence to raise HPC to the power of knowledge for machine understanding. This admittedly controversial but important presentation will assist the listener in distinguishing among emergent brain behavior, current interpolative and clustering putative machine learning, and Machine Intelligence for real machine understanding and the revolution that will follow.
