Human brain inspires new era of Big Data processing
Advances in ICT have transformed our economic and social lives. We are now able to bring distant people together in an instant, remotely coordinate disaster response, implement new medical diagnoses and treatments and achieve more accurate weather predictions and climate modelling. While all this – and much more – has been made possible by sustained computer performance improvements, standard processors simply cannot keep up with the exponential growth of data. This has pushed demand for the development of new fast, compact, low consumption processors. Applying brainpower ‘New processing and computing methods and storage devices are needed to tackle the increasing gap between data growth and processing power,’ explains FTK project coordinator Prof. Chiara Roda from the University of Pisa in Italy. ‘Our approach to this has been to develop image processing technology that emulates the architecture of the brain, which we then tested at the Large Hadron Collider (the world’s largest and most powerful particle accelerator) at CERN near Geneva in order to solve a specific Big Data problem.’ Some of the most interesting subatomic processes generated by Large Hadron Collider particle collisions are very rare and lie hidden within an extremely large amount of irrelevant events. Selecting interesting events from this ‘background noise’ in real time is therefore essential in order to fully exploit the potential of these experiments. The project team built “accelerators” for algorithms that usually take up a large amount of processing time and resources. ‘Our technology works by filtering significant information for further processing out of images that are too complex to be processed directly by standard computers,’ explains Roda. ‘This is how the brain processes images. For higher-level processing and long-term storage, it only selects data that matches a particular set of memorised patterns. Our technology emulates this low level brain function.’ Unlocking Big Data potential This capability opens up a large number of other potential applications ranging from astrophysical and meteorological calculations to robotic automation and security applications. ‘Communication with potential end users in the high energy physics environment was very good,’ says Roda. ‘ATLAS, an experiment at CERN designed to uncover the secrets of the universe, approved our technology to carry out a huge amount of online processing, which was initially expected to be done with commercial central processing units (CPUs). Outside of high energy physics, our objective was to accelerate and automate data processing for medical diagnosis.’ Indeed, the FTK project consortium has also been in discussions with companies interested in using the technology for smart camera applications and medical imaging. A young researcher who was awarded a Marie Curie scholarship to work in Pisa on “Magnetic resonance fingerprinting” (MRF) is studying the possibility of applying this technology in order to speed up algorithms based on pattern matching. ‘By increasing computing efficiency and power, we have shown that this technology can bring advantages to many areas of society,’ says Roda. ‘It can also save energy and space, and is less expensive than large CPU farms. I think that the smart use of combined technologies should be able to solve the Data Deluge problem that we all face, without the need for a huge amount of resources.’
Keywords
Large Hadron Collider, FTK, data, brain, ICT, climate modelling
