Research being applied to solve Big Data problems
The ACVT (now AIML) has been awarded two contracts to work with the Defence Systems Innovation Centre (DSIC) in the area of Big Data.
Big Data Intelligence Platform: Defence and Intelligence agencies are faced with an ever-increasing amount of incoming data. The ability to extract actionable intelligence from this data, whilst using limited manpower, will require new solutions to deal with the information overload. Big Data platforms such as Apache Hadoop are being used by internet-scale companies such as Facebook and LinkedIn to deal with this problem. They are becoming increasingly mature and can be applied to similar problems faced by Defence and Intelligence agencies. The ACVT is working with DSIC and UNSW to develop entity and relationship extraction techniques to solve these “Big Data” problems. Importantly, these techniques must operate at massive scale to be of use.
Automated Production of Geospatial Information: Defence and Security forces the world over are expecting a significant increase in the volume of ISR data over the foreseeable future. To be useful, the increasing volume of ISR data must still be processed, exploited and disseminated. This must be achieved with the same or reducing workforce. To achieve this it necessitates the increased use of automation in the analytical workflow. This research project will address this issue be developing approaches for a machine learning system that can be trained to extract and classify geospatial features from satellite and aerial imagery. The key objective of the project is to develop techniques to automate feature extraction and classification from satellite and aerial imagery based on a supervised machine learning system.