ARLINGTON, Va.—Developing interpretable data mining, machine learning and deep learning algorithms—as well as designing systems and interfaces—to enable novel ways of human‐machine interactions, including an improved understanding of challenges such as trust, explainability and resilience that improve human‐ autonomy partnership, is the work being by researchers at the University of New South Wales (UNSW).
Thanks to funding provided by the Office of Naval Research (ONR) Global, the team of scientists is boosting the capability of human‐machine interactions, which serve as the core of interactive intelligent systems, by exploring a new line of research based in future intent prediction, whereas most existing research focus on detecting, rather than predicting, intent. Future intent prediction is crucial in real‐life scenarios, where anticipatory response is required such as active sensing and autonomous navigation to make responses actively.
The ability of human beings to accurately recognize others’ intents is a significant mental activity that involves reasoning about objectives, such as what other people are doing, why they are doing it and what they will do next. Therefore, the quality of interpersonal and human‐machine communications can be enhanced through employing predictive intent analysis to identify other human beings as peers, competitors or bystanders, and to forecast their future activities. In the meantime, the situated courses of actions could be employed.
ONR Global Melbourne Office Science Director, Dr. Yoko Furukawa, believes that “collaborating with machines will become an essential part of how humans live, work, learn and play in the near future, as intelligent machines have the potential to seamlessly augment and boost humans’ physical and cognitive capabilities. For machines to become our collaborators rather than mere tools in a wide range of settings, however, many science and technology breakthroughs still need to occur. ONR Global believes that many of these future breakthroughs will come from the international research communities, and also recognizes that Australian research communities are particularly advanced in the field of human‐machine interaction science”.
Present and future applicability
Previous research has revealed that human intents could be inferred by measuring human multi‐ faceted activities from multiple heterogeneous information sources, such as body and brain sensors. These are sensors for detecting physiological signals like an electrocardiogram; brain signals like an electroencephalogram; and gestures and eye movement and voice from inertial measurement unit sensors like accelerometers, gyros, etc.
Nevertheless, many of the existing studies focus on detecting, rather than predicting, the intent, or recognize the intent’s predefined variables. The intent prediction has not yet been fully explored when only partial observations or few clues for the intent have been observed.
Paul Maddison, director of UNSW Defence Research, states, “There is perhaps nothing more important today in defence research than human-machine interaction. Capabilities that were until recently confined to the pages of science fiction have leapt from laboratories and into the field. Soldiers, sailors and air personnel are humans enabled and limited by their physical and cognitive capabilities.”
He adds, “Machines and AI algorithms can bring speed and clarity to the decision-making process across the entire chain of command, from the infantry soldier to the Task Force Commander.”
With theoretical foundations and a data‐efficient intent prediction paradigm, this project aims to boost the capability of human‐machine interactions, which serve as the core of interactive intelligent systems (e.g., robot, logistics units or other sophisticated military systems). Such systems aim to meet the increasingly complicated defense demands—namely, improve operators’ performance and training techniques, autonomous weapons, reduced casualty, post‐casualty recovery and mental health management.
For instance, future intent prediction could help autonomous vehicles to decide how to maneuver depending on the next predicted intent, or assist robots to make future decisions. In these scenarios, existing systems can only detect the intent when it has already occurred or partially occurred, which may not give operators sufficient time to respond.
“This research will be a critical piece in a wide range of human‐machine collaboration applications in the near future not only in both defense but in civilian applications as well,” says Furukawa. “For example, intent prediction will help build a safer and more flexible manufacturing environment where rigid mass‐manufacturing assembly lines will be replaced by a team of humans and smart machines that can flexibly adapt to customers’ rapidly evolving needs.”
In healthcare, she adds, “it will enable technologies for machine‐assisted rehabilitation and elderly care. In training and learning, it can help build a recommender system to sort and arrange the materials in a way the learner can most effectively digest and understand throughout the training process.”
ONR Global searches the world for leading partners that support naval and revolutionary multidisciplinary research through active worldwide cooperation, to solve present and future U.S. Navy and Marine Corps needs.
“The UNSW partnership with ONR Global is extremely important, as it brings together the joint research priorities of the United States Navy and the Royal Australian Navy,” says Maddison. “Australia is a maritime nation, and many of the geostrategic regional drivers revolve around controlling or defending core interests at sea. That means that leading-edge, undersea-sensing technologies will be required in our ships, submarines and aircraft. UNSW remains ready to build on our relationship with ONR Global to accelerate the development of these disruptive battle ready technologies.”
ONR Global sponsors scientific efforts outside of the U.S., working with scientists and partners worldwide to discover and advance naval capabilities.