Leading British artificial intelligence (AI) company Benevolent AI has announced that it is the first company in Europe of any kind to begin using the world’s most advanced deep-learning AI Supercomputer – the DGX-1 from NVIDIA.
Benevolent.AI is a London based start-up, which was formed by the management team of another start-up, Proximagen, a drug discovery and development company, sold to Upsher-Smith laboratories in 2012 for $553 million.
The Company describe themselves as “focused on the harnessing the power of artificial intelligence for the advancement of scientific discovery and using technology as a force for good by applying it to real world problems”.
The DGX-1 will be used to “process and analyse the vast amount of highly fragmented scientific research information to discover meaningful, intelligible results that will enable faster medical breakthroughs.”
The supercomputer will dovetail with Benevolent.AI’s unique and proprietary Judgment Augmented Cognition System (JACS). JACS employs deep natural language processing, machine learning and AI to “formulate new, usable knowledge from complex scientific information.”
JACS “augments, accelerates and advances scientific discovery…by analysing millions of scientific articles and hundreds of medical databases – facilitating faster breakthrough and delivering the potential for more precise and targeted medicines.”
Says Co-founder and Director of Benevolent.AI, Ken Mulvany, “there are 10,000 updates per day on PubMed alone, presenting an impossible challenge to scientists to keep up with the rate at which new scientific knowledge is produced.”
“This new AI supercomputer will boost our processing power and accelerate the creation of new relationships amongst disparate information sources to yield faster scientific innovation.”
The DGX 1 will also be used to accelerate Benevolent AI’s chemical modelling capabilities and enable deeper integration of chemical structural activity relationships using the company’s own AI algorithms.
Benevolent.AI are based in London Kings Cross’ Knowledge Quarter; the company is split into 2 distinct units; “Benevolent Bio”, which was formerly known as Stratified Medical, and leverages tech in the pursuit of human health advancements and the exploration of further bioscience applications, and “benevolent technology” which refines the company technology and widens the scope of the products across applications and industries.
Besides Mulvany, Jackie Hunter PhD CBE FBPharmacolS FMedSc who looks after BenevolentBio as CEO, and Jérôme Pesenti, CEO of BenevolentTech, sit on the board of Directors of Benevolent AI.
3 co-founders, Michael Brennan, Ivan Griffin, and Brent Gutekunst, Head of Corporate Development, VP Business Development and Chief Development Officer respectively, comprise the Senior leadership team. The company are currently hiring.
Described as “The world’s first supercomputer in a box”, by maker NVIDIA, the DGX-1 is built with NVIDIA Pascal™- powered Tesla®P100 accelerators, interconnected with NVIDIA NVLink™. Pascal, says NVIDIA, “was designed as the engine of computers that learn, see and stimulate our world – a world with an infinite appetite for computing.”
The computer’s software stack comprises an NVIDIA Deep Learning SDK, the DIGITS™ GPU training system, drivers and CUDA® for “rapidly designing deep neural networks (DNN), as well as access to cloud management services.
Applications on the DX-1 will run up to 12x faster than with previous GPU-accelerated solutions, says NVIDIA, hence the reason Benevelont.AI are so pleased to have acquired one.
BigData and “machine learning algorithms” have been in vogue for some time, and given the ability of both to sort through more information than ever before and make better decisions than ever before the sector still has to be, and even more so in the light of success stories like Salesforce, or Zuora, or even Facebook, for that matter, probably still the hottest sector for disruptive start-up activity.
According to Crunch Base in August 2015, Benevolent.AI raised $87.7m of venture capital funding, from Lansdowne Partners, Lundbeck, Upsher Smith Laboratories and Woodford Investment Management at an undisclosed valuation.