In 1907 German chemists at Henkel AG combined sodium silicate with sodium perborate, which when mixed with boiling water resulted in a self-acting textile-friendly and odorless bleach they eventually branded Persil. Nearly 100 years later Persil remains the detergent of choice for millions of households, manufactured and sold globally by multi-nationals including Unilever and Henkel.
Those same chemists could never have imagined that decades later their colleagues would be using data-centric cognitive computing and advanced analytical simulations to improve upon their formula.
“It’s the perfect machine learning project. Vegetable oils are one of the most challenging stains to remove from textiles. Using machine learning and high performance computing, IBM researchers are working with Unilever and researchers at the Hartree Centre in Daresbury, UK, to calculate the various phase changes of the chemicals. As more simulations are done, the system learns and goes on to select the next one,” said David Watson, director of the IBM Research team based at the Hartree Centre.
The research is part of a collaboration with the Science and Technology Facilities Council’s Hartree Centre and IBM, which launched a little more than a year ago as part of a five-year project with technology and onsite expertise worth up to £200 million, including data-centric cognitive computing technologies, such as IBM Watson and the latest OpenPower hardware with some special technical tuning from IBM’s R&D labs.
Alison Kennedy, Hartree Centre’s director says, “This collaboration between Hartree and IBM provides a fantastic and unique opportunity for the UK to tackle a wide range of industry-led challenges using a range of novel and emerging technologies. We are the only people drawing on the combined power of computer simulation, large scale data analytics and cognitive computing to offer solutions to industry, to enable them to design better products, and bring them to market more quickly in a cost-effective way. The combination of world-class scientific research, state of the art technology and skilled personnel provides a winning combination that delivers competitive advantage to our industry partners and offers new insights into the interplay between research and innovation.”
Watson adds, “We took a careful look at what scientific challenges could be addressed in the short term to create a long term economic impact in the UK.”
Continuing the collaboration’s progress
Working with industry in the UK, the Hartree Centre team has narrowed its data-centric computing projects to focus on four critical economic growth areas important for the country including chemistry, life sciences, enabling technologies, and engineering and manufacturing.
IBM scientist Katharina Reusch, who studied neuroscience and electrical engineering at the University of Nottingham, recently joined the IBM Research and Hartree Centre collaboration which now has 10 IBMers. She is working on the enabling technology research within several sub-areas, including large scale machine learning, Internet of Things, supercomputing in the cloud and visual analytics, among others.
“We are at an incredible inflection point due to all the data and the compute power now available. I am excited to begin working with colleagues in the UK, as well as our labs across the globe to apply these tools to solve industry’s challenges,” Reusch said.
Another member of the team working at the lab is 30-year IBM veteran Evan Grant who studied metallurgy and materials science at the University of Cambridge. “The ability to apply IBM’s technology directly to the challenges of the UK through the Hartree Centre is very motivating to me,” said Grant. He cites many goals over the next four years:
“With Watson Internet of Things and maintenance data we can improve the efficiency of production lines. In addition, our OpenPower computers will model everything from mechanical parts to designing new drugs. A few years ago this would have been a wish list, but today it’s all possible.”
“A lot of what we are trying to do is motivated by industry applications and workflows driving the design of IBM’s future hardware systems. This research is providing additional diversity of how industry can use these applications, which is exciting for us at IBM Research because science without application doesn’t lead to game changing innovation,” said IBM Distinguished Engineer and Chief Science Officer Kirk Jordan of IBM Research UK at the Centre.
Join the team
Both IBM and the Hartree Centre are actively hiring regular and post-doc scientists with skills in deep machine learning, high performance computing, applied mathematics, computer science and computational science and engineering. (Current job openings: Geospatial Agriculture research Staff Member, Research Staff Member Chemistry, Research Staff Member Acceleration of Big Data/Machine Learning, Research Staff Member for Data Centric Cognitive Systems, Research Staff Member Graphics and Visualisation project, Research Staff Member Containers as a Service, Post-Doc Researchers)
Watson again: “The strategy is in place, the technology is installed, and we are continuing to actively hire talented people to join our current team working on these exciting projects.”