DailyPost 845
DEEP RESEARCH
Even the methodology of research has no end, leave aside it being enormously multidisciplinary. Vistas of research is defined by the imagination of the researcher, his expertise, effort and the endless perseverance he ought to possess. Minds and paths will open up and take human life to a different level. Two of the co-founders of DeepMind, Demis Hassabis and Shane Leg are the epitome of this research philosophy. Unfortunately, this vision most of the researchers don’t share, it is grabbing of product called thesis, for self advancement.
UCL’s computational neuroscience unit is where it all started, Hassabis has a PhD in the subject. The basic strategy of “developing AI by teaching computers to mimic the thought processes of our own brains. The main learning was how to use information to make decisions and learn from our mistakes. British AI start up DeepMind, relatively unknown, was bought by Google for $600 million in 2014. This provided the required impetus and it continued to refine it’s neural network driven technology. It is taking machine learning to an amazing trajectory, particularly deep learning.
AlphaGo has been it’s iconic success. In a project in collaboration with London’s Moorfield’s Eye Hospital, DeepMind has been given access one million images from historical eye scans and patient data. It’s success would mean spotting early signs of degenerative eye conditions much earlier than a human can. Another project in collaboration with the UCL’s radiotherapy department aims at reducing time to plan treatments. In some complicated procedures it has been brought down from four hours to around one hour. Streams, an early warning app would start from detecting acute kidney injury to potentially learning to spot many other conditions is used by NHS.
Research need not get completely connected to profits. Being true to core belief that knowledge – or data – is the real prize. The refinement of cognitive processes, enabling to tackle complex problems would be a boon for mankind and for sure will yield immense commerical benefits. Short term financial gain is the antithesis of research. Gains will happen in unknown areas. Using this technology, power consumption needed for cooling at Google’s massive data centers has come down by 40%.
OUR HEALTH DEPENDS ON THE HEALTH OF AI.
Sanjay Sahay