|2017-08 Research Impacts and Machine Learning|
Compucon CPD Seminars
Near Term PC Technologies and Research Impacts
As Moore’s Law is hitting the wall, divergence of computing hardware designs has emerged and the previous market views on computers being commodity will lose support fast over time. We are among the first to spot the change and to invest in researching the change. Two years of efforts have been consolidated into an On-Ramp folder as an illustration of how one system can perform 5 selected open source applications from 5 to 200 times faster than another similarly priced mainstream system. This folder did not receive any echo within our peer group but the insights, when presented in an open seminar in AUT recently (which was attended by over 40 persons including 5 of our peers), were rated highly by a couple of senior academic people afterward. The insights were not just technical but structural as well. This session will repeat some of the key slides from that seminar. Building further on the output of our research efforts, we have been able to carry out another investigation for the international SKA project and provided some contributions that SKA would benefit. The info is relevant to our peers and some info will be repeated in this session as well. For peers who plan to stay on in the industry for the next 5 years or longer, this session is strictly essential for business positioning. For the immediate future, this session will preview the resumption of a Compucon system platform that has submerged for 10 years and it will come back soon in a big bang. This session will mention SSD as well because SSD has replaced HDD as the preferred storage medium in desktop and high performance computers.
Machine Learning Business Potentials
Machine learning is another research and development agenda that has been mentioned regularly for about a year by now. The last occasion was a seminar (attended by 8 peers) in which a university student explained what he did to train a software application to identify hand-written digits with 99% accuracy. The example is just the beginning. We plan to prepare 5 models as our second on-ramp folder for next March release. The sampler is to allow our peers to have hands-on experience of what machine learning can do in the real world and especially what is accessible by SME (large businesses know it already). David will provide his views on the business potentials of machine learning or if there is any at all in this session.