The 2020 TIDE Hackathon took place at the IBM Watson Centre in Munich, 17-21 February 2020.

TIDE Hackathon is an event within the NATO’s Allied Command Transformation-led Interoperability Continuum (the others being TIDE Sprint and Coalition Warrior Interoperability eXercise), and provides an opportunity to experiment with interoperability challenges from emerging and disruptive technologies.

The TIDE Hackathon encourages unconventional approaches from people with fresh ideas; it encourages energetic enthusiasts that are not constrained by what they think we want to see or hear. In a nutshell, Tide Hackathon is innovation in practice.

The IBM Watson Centre is a market leader in digital transformation with industry expertise in intelligent assets (Internet of Things), data, artificial intelligence, cloud, security, block-chain and quantum computing: securing this high-tech venue, by collaborating with IBM, provided a fitting backdrop to the 2020 TIDE Hackathon as NATO challenged teams from the military, industry and academia to develop machine-learning solutions that will improve command and control interoperability in support of Allied Command Transformation’s Warfighting agenda.

Why Machine Learning. Machine learning is a key component of data driven decision-making and is fundamental to the success of data driven organizations. Through modelling and knowledge capture, it vastly reduces the amount of effort required by staff to complete routine tasks. It does not replace human judgement; instead, it captures human understanding in the form of rules, data models, mathematical models and interactions between people and data. An example of machine learning in NATO’s Allied Command Transformation is “AI-Felix”, an innovative approach that uses algorithms to sort and re-direct documents as they enter our building, saving hundreds hours of manual work. This is a very specific example; in the future we will see fully interoperable machine learning solutions enhance decision making across multiple functional areas and between mission partners, whether they are NATO nations, partner nations or other organisations.

Machine learning within the context of emerging and disruptive technologies was discussed at the Fall 2019 TIDE Sprint. This event provided the context for the 2020 TIDE Hackathon and the following machine learning use-cases that teams were asked to solve:

Dynamic Labelling of Voice and Video. Remote video and voice conferencing is an increasingly important and mission-critical operational service, both in peacetime and for operations. As NATO Commanders work more closely with partners and non-military actors, it will be important for them to switch the classification of an audio stream mid-conversation, in order to control in real time who has access to sensitive information. This challenge requires teams to identify the security classification of teleconference content, using machine models to limit classified content to appropriately cleared teleconference participants in real-time.

Extraction of Structured Data from Documents. NATO has a repository of unstructured and unclassified documents including doctrine, directives, standards and specifications. Information products are derived from these documents, but the volume and unstructured nature of documents, means that to understand and extract information products from this wealth of documentation is beyond purely human resources. This challenge requires teams to use Natural Language Processing, or NLP, to find entities of a certain characteristic, or profile in the text that can identify information products and improve our understanding of how Command and Control processes are executed through the exchange of information. Successful solutions could contribute to NATO’s Command, Control and Communication Taxonomy and development of future interoperable systems.

Predicting Crisis. Crisis detection in a complex environment requires a holistic set of indicators, that nationally or regionally identify developments or trends that may spark a crisis. However, global indicators often lead to globally good comprehensive models that sacrifice regional indicators. This challenge requires teams to use machine learning to detect and monitor regional phenomena such as the price of goods, crime rates and regional poverty, in order to predict a looming crisis. IBM raises this use-case as an advocate of military analysts, and an example of where humans can be assisted in the analysis of massive amounts of data. This is in line with the IBM’s strategy on Artificial Intelligence principles, which state that human intelligence be augmented and amplified by the use of Artificial Intelligence.

Twenty teams provided solutions to the challenges, and the winner was ‘Team FRONT’ - a team of Software Developers from the Polish Ministry of Defence IT Projects Centre. They demonstrated how, using open source tools, it is possible to restrict access in real-time to live streamed videoconference content according to the security clearances of participants. This is an important contribution to the Data Centric Security Vision and Road-Map and demonstrates how NATO could communicate with mission partners  in a complex environment of multiple network classifications. Team FRONT will be sponsored by ACT to attend the Spring 2020 TIDE Sprint.

Every TIDE Hackathon deepens our understanding of the application of emerging information technology and its potential support of and impact on interoperability.