Saturday, January 5, 2013

Research computer

Recent work by Emory mathematician, Ken Ono, finds further explanation and meaning in the work of Ramanujan, who died nearly a century ago. Applications stretch in many directions including explaining some behavior of certain types of black holes. This illustrates that ground breaking insights originate from pure imagination even in the absence of a formal education and proven tools. More importantly, it is possible that systematization of education, tools and processes is at the heart of declining productivity in true insight generation.

Most of today’s science is geared toward proving a stated hypothesis. This is akin to manufacturing – execution of a defined and specified process, that includes design of experiments, data gathering, statistical analysis and hypothesis testing. Indeed, humanity is fast approaching the “singularity,” in which computers can be trained to conduct research better than themselves, for the research process is the most systematized and computer programmable than anything else. It is possible that the first type of Artificial Intelligence that may become practical is a research computer, able to design studies, collect data, analyze, test and prove a hypothesis given to it. Such a machine, however, will be incapable of imagination and insight generation. So, it may walk, move objects and conduct research, but it will never sleep, weep or dream.

Perhaps, the “singularity,” that many look forward to is related to a complete elimination of imagination from the human psyche.