Vanguard Magazine

T TEChNOlOgy WATCh 38 JUNE/JULY 2017 www.vanguardcanada.com full-time employees in Sherbrooke, includ- ing artificial intelligence research scientists and developers. Menya Solutions recently launched Hy- bridLogic, a suite of artificial intelligent algorithms designed to help humans, ro- bots and drones analyze a tactical situa- tion and make timely decisions during defence operations. This technology in- cludes intent, capability and opportunity analysis algorithms that process sensor data, human input, and a prior knowledge to predict threats, infer their goals and plans, and understand their vulnerabilities and the risks they pose to protected assets. HybridLogic also includes algorithms that generate a course of actions against recog- nized threats and support the execution of these actions by monitoring contin- gencies and revising the plan accordingly. The algorithms solve a variety of situa- tion-understanding and decision-making problems to provide artificial intelligence capabilities to agents (humans, robots or drones), operating alone or in multiple coordinated units. The algorithms take into account various constraints related to defence operations in particular, including spatial constraints, communication con- straints, computational constraints, and resource constraints. "Having humans collaborating with robots or drones in defence operations is no longer a science fiction concept," said Froduald Kabanza, CEO of Menya. "The ability of drones to interact naturally with humans and to comprehend their environment will largely depend on their proficiency to recognize the goals and plans of other agents – a proficiency that technologies like HybridLogic are meant to provide." Kabanza went on to explain that, on the other hand, the unprecedented complex- ity level of defence operations brought by these technology developments is such that humans can no longer control every- thing without the aid of artificial intelli- gence algorithms at various levels. "Com- panies developing command and control systems for defence operations can use HybridLogic algorithms to automate the analysis of threats and support the aware- ness of human operators to monitored situations," he said. Companies who are in the virtual reality training solutions area for defence operations can also integrate HybridLogic to present enriched training scenarios, according to Kabanza, "involv- ing simulated entities that can make au- tonomous decisions." I've always said that data is very pow- erful, but it always comes down to "so what," and using data as a lever to de- termine the answer to your problem and drive facts-based decision making. Another company making waves in AI is Motsai Research, an electronic hardware design company, which extracts meaning- ful information from miniature motion and environmental sensors. By combining the most advanced integrated electronic wireless technology and pushing the lim- its of sensor miniaturization, they create sensor solutions that acquire, process and classify motion patterns within the sensor itself, which incorporates custom elec- tronic hardware, advanced sensor fusion algorithms and assisted machine learning to create powerful information-generating systems. This enables an important reduc- tion in system complexity. Motsai is re- ally focusing in on supporting the defence sector in areas such as soldier protection, machine operating conditions monitoring and advanced warning of abnormal situ- ations. Novel miniaturization techniques, encapsulation methods and energy har- vesting solutions allow sensors to be inte- grated inside objects and systems in ways that were not possible only a few years ago, which is driving growth in the Inter- net of Things market. But the next step is making sense of the data and using it to drive machine learning, which is exactly what Motsai is doing. Sensors are at the frontier of cyber phys- ical systems and will provide better situ- ational awareness and improve decision- making abilities of future defence systems. For this potential to be realized, future sensors must provide meaningful informa- tion and be fully integrated within larger systems using secure and robust commu- nication mechanisms. Machine learning and artificial intelligence can help lever- age the potential of distributed sensors at many levels, and future systems will very likely integrate sensors in large scale de- ployments. On the academic side, McMaster Uni- versity has quite the history in the evo- lution of smart systems, including be- ing creators of cognitive approaches to processing data from diverse sensors and the development of efficient optimization techniques. McMaster is working today to leverage this strength and bring a smart- systems approach to a broad range of in- dustries within Canada in areas such as Nanoscale research, which uses machine and deep learning algorithms to provide quantitative insight into materials micro- structures. They are also using machine learning to help with weapon detection, building more intelligent vehicles, and healthcare. On the big data front, McMaster is working with IBM to improve big data quality by designing and building intuitive software tools to keep personal informa- tion private while automating this process for organizations, which saves them time, money and improves data analysis results. In the battlefield, they can apply big data to troubleshoot processing problems, and to collect and analyze data from a variety of sources, including temperature sensors, pressure-flow monitors, chemical-con- The size of the sensor modules Motsai Research uses to collect data, compared to a Canadian 10 cent coin.

• Cover