A University of Oxford spin-out firm Mind Foundry is aiming to revolutionise the use of machine learning as applied to business data with the launch of a platform which cuts out the need for data scientists.
The platform was unveiled this month at the Ashmolean Museum in Oxford as part of Mind Foundry’s commercial launch. The goal is to humanise machine learning.
The firm says the cloud-based platform allows anyone, of any technical ability and in any sized organisation to unlock the full value of ever increasing volumes of data and make decisions on complex business issues without the need for data scientists.
It contends that most commercial services on offer to date have been aimed at data scientists, requiring firms globally to compete for scarce talent.
Instead Mind Foundry says it platform gives problem owners within firms the bility to use machine learning to interrogate their own data using a system which compares with commonly and easily-used platforms such as Excel or data visualisation platform Tableau.
How Mind Foundry works for businesses in its own words
The platform automatically builds appropriate machine learning solutions for business problems in minutes or hours, rather than weeks or months, and provides full transparency and auditability of solutions. Problem owners are guided through simple steps to develop and deploy models. Along the way, the platform gives guidance and advice to enable continuous improvements, discovery of actionable insights and a complete understanding of machine learning derived results.
Mind Foundry was spun out of the University’s Machine Learning Research Group in 2015 by the firm’s founders and chief scientific officers Professors Stephen Roberts and Mike Osborne.
The firm, which is a portfolio company of the University of Oxford with lead investors Oxford Sciences Innovations and Parkwalk Advisors, recruited Paul Reader (pictured) as CEO in December 2018 to spearhead the platform launch.
Reader was previously chief strategy officer for digital marketing platform Wayin and global head of strategy and organisation development at Oracle. Before that he worked at Eloqua which was acquired by Oracle in 2012.
The platform was developed working with large investment firms, telecoms and manufacturing firms. Mind Foundry says it is applicable not just across the full range of sectors but also across a full range of business functions within firms.
Reader says: “We are developing machine learning solutions for functions and departments across enterprises such as marketing, sales, finance, R&D and services. In fact, we are integrating Mind Foundry across our own business across all the functions to show how any business of any size can answer real businesses challenges from their data easily.
“Our technology is the first of its kind to use AI to guide business users through an end to end machine learning solution development process in a way that augments their skills. We offer human intelligence amplified.”
Discussing the platform, director of engineering Rob Williams says: “There were tools out there, but they were aimed at the data scientists, so they were a different type of tool. We wanted a tool aimed at the problem owner – to give them a user experience which was like the tools they were used to such as Excel and Tableau.
“We want people to bring their problem and bring their data with no machine learning expertise required.”
Williams says one obvious use for the platform would be to predict which customers are likely to complain and even to put a numeric value on the likely cost of a complaint
“To make a prediction you need a model, to build a model you need historic data, records of a known outcome. Along with that you need as many facts as you can gather. We can train a model to learn the patterns between the facts and the outcome.”
Professor Roberts adds: “At a time when data science skills are in short supply, Mind Foundry is aiming to create 1,000 ‘citizen data scientists’ by 2020 and ensure humans are always ‘in the loop’ when it comes to machine learning.”