Following the largest ever competition run by Defence Science and Technology Laboratory’s Centre for Defence Enterprise (CDE), Massive Analytic Limited, an artificial intelligence firm, has been awarded a contract with the Ministry of Defence (MOD) to develop technologies for ‘autonomy and big data in defence’.
CDE funds novel, high-risk, high-potential-benefit research to develop cost-effective capabilities for UK armed forces and national security.
The competition called for a proof-of-concept research proposal for innovative component technologies and techniques to acquire, process, analyse and visualise data to support decision making by both humans and autonomous systems.
The big data analytics platform
In its winning submission, London-based Massive Analytic proposed a big data analytics platform for operational tactical planning and support, using persistent surveillance and its own patented artificial precognition, the brain behind its analytics platform, Oscar.
The system will bring together multiple static and dynamic surveillance data sources, outputting predictions with high degrees of accuracy. When completed, the project is intended to mitigate the effects of data deluge on military surveillance analysts by providing advanced autonomous functions, facilitating the dissemination of critical insight and intelligence across the chain of command via an interactive data dashboard.
The platform has been designed for non-combatant evacuation operations, to support military teams in ‘cold scenarios’ – where they might not have much real information on the ground. Or whatever information exists, is spread across different systems and can take time to relay to ground troops.
Massive Analytics’ big data analytics platform would enable big chunks of information to be quickly disseminated amongst key players, all on a secure system.
Analytics enhances data value
Live insights from real-time data streams will become clearer and actionable far more quickly allowing quicker responses to immediate threats. Advanced planning for a diverse range of peace keeping and humanitarian scenarios will become sharper, thereby mitigating risks and allowing planning for possible eventualities.
Underpinning this new platform are several innovations including video precognition (a type of deep learning based on Oscar’s evolving artificial precognition) and multi-stream data synthesis.
The system is able to integrate video feeds to identify behaviours, such as hostility, allowing for advanced planning for diverse and evolving scenarios. The system also boasts over 99% accuracy in facial recognition.
Software testing strategy
Speaking exclusively with softwaretestingnews.co.uk, Gary Brooks, Chief of Engineering at Massive Analytic explained the testing strategy behind the big data analytics platform.
“For the main analysis part of the platform, we carry out a great deal of unit testing using Jenkins to ensure that unit tests are automatically run as updates are continuously deployed into the development environment,” Brooks said. “We work in a continuous deployment, agile fashion utilising three environments. Unit testing is carried out on a nightly basis, ensuring testing is not a hindrance to delivery.”
Brooks also explained that for the UI components, unit testing was not sufficient due to the rapid changes in requirements.
“For the front end we’re using primarily offshore testing through partners,” he said.
While the big data analytics platform was developed for defence and security purposes, a new market in video advertising has appeared.
“We’ve had request from digital media firms interested in analytics breaking down what makes a good video versus a ‘bad’ one,” Brooks said.
Cat videos and ground troops may very well be supported by the same analysis in the future.
Written by Cecilia Rehn