LONDON, UK - May 20, 2015 - Massive Analytic, a start-up specialising in innovative precognitive enterprise analytics, and the Department of Computer Science at University College London (UCL) have secured a 12 month Knowledge Transfer Partnership (KTP) to collaborate on the patented technology artificial precognition that drives the Oscar data analytics platform. The initiative will bring improved and ground-breaking predictive and automated data discovery capabilities from academic research and into the service of enterprise analytics.
The co-operative venture, co-funded by Innovate UK the UK's innovation agency (formerly the Technology Strategy Board), will see the development of intelligent recommendation algorithms and creation of a controllable prediction engine that will add even more powerful artificial intelligence (AI) capabilities to artificial precognition - the "genius" behind Massive Analytic's Oscar platform. The project will address the challenges of fully automating data science workflow and achieving scalability. The goal is to be able to predict outcomes with the absolute minimum human input. For example, if you tell Oscar you want to predict revenue, it will select the processes and algorithms completely autonomously to make predictions, learning continuously from the actual outcomes to make even more accurate predictions over time.
The team, led by Dr. Jun Wang and Dr. Tamas Jambor, joined by Dr. Kleanthis Malialis, formerly a postdoctoral student at University of York, will harness academic capabilities in intelligent recommender systems to develop novel uses of optimal control theory for real-world applications.
Dr. Wang, the academic supervisor for UCL Computer Science, said, "The KTP project will explore practical methods of automating the selection of machine learning algorithms for big data analytics. We are excited about the KTP project with Massive Analytic as this provides a channel for us to transfer our research results in information retrieval and recommender system at UCL into actual products and to be directly used by end users."
Optimal control theory is a mathematical framework to streamline decision-making processes, such as those involved in improving customer satisfaction - a far superior version of the simplistic algorithms applied by retail websites to make customer recommendations based on learned preferences and purchasing decisions made by other customers. By embedding optimal control algorithms in Oscar, proactive targeting of customers will become more accurate and have greater relevance. Feedback from the success or failure rates of these recommendations helps Oscar to modify its behaviour, become more expert, and improve customer satisfaction.
The results of the current project will enable Massive Analytic to launch a fully commercialised Artificial Precognition (AP) solution which not only improves current data analytics systems but gets better as it learns and creates its own "tricks of the trade".
George Frangou, the Executive Chairman of Massive Analytic, said, "The KTP with UCL will help develop Oscar into generalized application for emulating a user's thought processes, through the use of optimal control theory and the development of a universal recommendations for business algorithms. Business users will be able to easily carry out advanced analytics and data scientists will have time freed up to discover more insights. The novel use of optimal control theory will optimize processes to achieve the required output, learning on the fly from successes and failures."
The partnership, supported by UCL Advances, will develop a universal selector engine for business analytics algorithms embedded within Artificial Precognition (AP), leading to a radical innovation, a solution to the need to harness machine inferential thinking in diverse emerging applications. It will recommend the best way of approaching a particular problem taking into account the entire analytics workflow. This will save time for data scientists and help business users reach better insights more easily.
As a highly innovative software company, Massive Analytic has set its course to develop world-leading data solutions based on its patented Artificial Precognition (AP). Oscar, its first platform to use innovative artificial intelligence (AI), emulates the characteristics of human thinking to understand business data and make predictions. The next generation, based on the KTP project, will enable AP to function as an autonomous solution to emulate human deductive powers, thoughts and processes to enable intuitive leaps to be made when exploring and analysing big data.