Next Generation Applied Artificial Intelligence
Montvieux receives DASA funding to enhance the protection of forces and improving the efficiency and effectiveness of information collection
Challenge
The proliferation of data within the military poses a significant challenge for operators interpreting differing data sets into meaningful information upon which to make informed and timely decisions. Artificial Intelligence capabilities are developing at pace and can present opportunities through Deep Learning for operators and decision makers to interpret vast, disparate data sets concurrently.
Solution
The Defence and Security Accelerator (DASA) has funded two projects led by Montvieux, in excess of £500,000, following a DASA themed competition to find new technologies, processes and ideas to ‘Revolutionise the human information relationship for Defence’.
Prediction Toolset
The Prediction Toolset is a Deep Learning based Artificial Intelligence capability that uses current and historical information to predict the change of control on the ground, in both space and time, between opposing groups fighting within an operational theatre.
This capability provides foresight to analysts and collection managers, enabling them to proactively anticipate future events on the ground, thereby enhancing the protection of forces and improving the efficiency and effectiveness of information collection.
Machine Learning Analytics as a Service (MLAaaS)
Preparing training datasets for Machine Learning algorithms is a resource-intensive task. In many cases, there may not be sufficient real-world examples of an object or scene to create an adequate labelled training dataset from. The impact is that automated detection tools cannot be used and analysts have to revert to manually reviewing images for important objects of interest.
The MLAaaS capability uses a Generative Adversarial Network (GAN) based Artificial Intelligence approach to the generation of labelled training data for use in 3rd party imagery object detection and classification systems. The solution creates user-defined scenes within a 3D synthetic environment, then using specialised GAN techniques, refines this image to appear real, before labelling features of interest, ready for use as a training dataset in existing Machine Learning tools.
This approach is able to automate the generation of realistic training datasets at scale including controllable scenes featuring objects that are infrequently observed in live data.
This capability will save imagery analysts considerable time, both in preparing training datasets, but also in manually reviewing imagery, as their existing tools can now be effectively trained to recognise rare or unusual, but significantly important features and objects.
Pete Webb, Managing Director at Montvieux: > With the help of DASA, we have been able to plan-back from potential commercial outcomes and identify the right industry partners to move this forward.
Benefits
The Prediction Toolset has significant applications across the Defence and National Security domain, including predicting disorder, financial fraud and aspects or organised crime funding. The technique could be successfully applied to predicting the fluctuations in finance markets (FOREX), location-based demands for good and services, predicting utility demands, or supporting improved prospecting for natural resources.
Finding or generating appropriate datasets is key to all areas of Machine Learning. The MLAaaS technique could provide a rapid mechanism for generating tailored datasets for those organisations involved with the development of optical systems or analysis tools. Interestingly the technique could also be applied to the generation of training data for different domains, such as acoustic or sub-surface acoustic (sonar).
MLAaaS could save many hundreds of engineering hours currently used to manually generate training data. This approach would enable organisations to more rapidly develop, test and optimise Machine Learning solution, accelerating AI-based research and development activities.
Next Step
Montvieux would like to further identify Prime / Tier 1 partners with whom they could inject / license the technology for exploitation within the Prime’s existing toolsets and capabilities. Montvieux is also keen to discuss the potential for the establishment of a dedicated spin-off venture, focussed on exploitation of their technology to address specific problems in non-defence market areas.
Contact
John Barrass, Chairman: john.barrass@montvieux.com