07/12/2017

Digital Catapult launches Machine Intelligence Garage

Author: Digital Catapult Staff Company: Digital Catapult News

Programme set to supercharge artificial intelligence development, supporting industry ambition to contribute £232 billion to the UK economy by 2030

London, 7th December 2017: Today, Digital Catapult launches Machine Intelligence Garage, a programme designed to help companies drive machine learning (ML) and artificial intelligence (AI) development in the UK.

The initiative, previewed in last week’s Industrial Strategy, will support the government’s ambition to make the UK a global centre for AI development, an industry anticipated to add £232 billion to the UK economy by 2030.

Set to commence in January 2018, Machine Intelligence Garage is a programme designed to empower AI and ML businesses, by providing access to computation power and expertise. Access to these vital factors is a significant challenge for start-ups in this area. A recent Digital Catapult study, surveying approximately 10% of the UK’s AI and Machine Learning start-ups, found that over half are constrained in their growth by access to computation power.

Dame Wendy Hall comments: “As recognised by the recent “Growing the artificial intelligence industry in the UK” independent review by Benevolent.ai CEO Jerome Pesenti and myself, access to computation power is amongst the significant barriers to innovation around Machine Learning and AI. It is great to see Digital Catapult’s Machine Intelligence Garage programme addressing this barrier for start-ups, and we hope to see cross fertilisation with other organisations such as the Alan Turing Institute that will be looking into similar challenges for the academic community.”

The next generation of “deep learning” approaches have been reaching human-level performance across a variety of domains recently, powered by very large data sets. However, these techniques incur a correspondingly large computational cost. This forms a barrier for UK innovators and researchers, where a single training run for a ML system can cost upwards of £10,000.

At the same time, significant innovation is occurring in computer architectures for machine learning. Besides established large players releasing new systems, such as the Amazon EC2 P3 instance, Google’s “Tensor Processing Units”, or Apple’s “Bionic” chip, there’s also a wave of innovative UK companies making incredible strides in this area. Exemplified by Bristol-based Graphcore, these are building specialised AI hardware – with recent funding rounds of US$110M in the past 18 months.

“The UK has all the right ingredients to become a global leader in applying artificial intelligence and machine learning,” comments Dr. Marko Balabanovic, Chief Technology Officer at Digital Catapult. “We have the academic strength, the ambitious entrepreneurs and great success stories already. With innovation in computing power and hardware for AI moving at unprecedented pace, Machine Intelligence Garage will provide UK businesses applying AI the support, expertise and compute resource they need to reach their full potential.”

To create Machine Intelligence Garage, Digital Catapult is collaborating with hardware providers, cloud computing providers, high performance computing facilities and academics. These include Amazon Web Services (AWS), Google Cloud Platform, NVIDIA, Graphcore, STFC Hartree Centre, EPCC, Newcastle University, the Alan Turing Institute, Bart’s Health Trust and Capital Enterprise. The programme’s initial funders are InnovateUK and ERDF.

In addition to facilitating access to computation, Machine Intelligence Garage will host a series of investor pitch days, experimentation days, meetups and showcases. These types of activities enable companies to grow faster, while learning from one another.

The programme is available to companies that are developing products or services that use ML or AI. Applications will be assessed based on the strength of the idea submitted and technical implementation plan, availability of data, and the immediacy of the need for computation power. Companies’ ethical use of data will also be paramount.

Machine Intelligence Garage will run for three years. Applications for the programme will open every six weeks, with the first Open Call going live today. The initial group of successful companies will join the programme on 23rd January 2018.

The programme is delivered by Digital Catapult as part of the CAP-AI project and is part funded through the European Regional Development Fund and an Innovate UK grant.

To find out more information and stay up to date please visit www.migarage.ai