Prof. Kaminka’s Lab

Prof. Gal Kaminka, an Associate Professor at the Department of Computer Science and a member of the Gonda Multidisciplinary Brain Research Center, heads the MAVERICK group (Multiple Agents, Virtual Environments, Robot Intelligence, Cooperation and Knowledge) at the Department of Computer Science.

Robots are autistic; humans are not. Kaminka and his MAVERICK group study social intelligence, especially as it relates to robots, and agents in virtual environments.

They are interested in building socially intelligent robots that are able to reason, manipulate, collaborate, and coordinate with other robots and humans.

The overarching goal is to develop computational models that explain social intelligence, allow for its replication and facilitate behavioral predictions.

Kaminka and his team view the wide gap between robots and humans as a computational challenge: even given the right mechanics and electronics, robots cannot interact effectively with others, unless their computational brains are made to think in a particular way.

The group’s research in multi-agent and multi-robot systems has two complementary themes: the control of teams of robots and agents on the one hand, and its complementary monitoring by the agents themselves on the other.

The two themes are inseparable: one cannot build an effective controller without having a feedback mechanism that monitors the results of the controller’s actions. In multi-robot systems, this translates to endowing robots with the ability to take social actions, and to understand those of its peers.

One of the seminal results of Kaminka’s research is the realization that teamwork, as a set of general mechanisms for collaboration, can be automated and computerized. This makes it possible for robots to collaborate well, cheaply, and do so with humans.

Applications of the group’s research cover a wide span: their work on multi-robot patrolling, on multi-robot exploration, and robot formations (convoys) was highly publicized, and resulted in several technology transfer and patent programs.

In addition, their work on intent recognition led to efficient, novel approaches for detecting anomalies and suspicious behavior. Work in distributed multi-agent systems led to a number of patents in mobile devices.

Much of Kaminka’s MAVERICK research involves communications, physical security and cyber-security. The MAVERICK group’s research into advanced models for cooperation within teams of autonomous robots and agents uses—and guides the use of—communications to support coordination.

The group’s research in strategic adversarial patrolling and exploration of buildings targets specific defense and homeland security applications. In particular, Kaminka’s group is one of very few that addresses the challenges of a single human operator controlling multiple robots in security tasks.

In another project, Kaminka and his group have developed the world’s fastest algorithm for detecting anomalous—and possibly suspicious—sequences of action, a technology that can be applied to automatic video surveillance and cyber-security. Kaminka is also building computer-based cognitive models that can understand human actions and intent.

The MAVERICK Group continues its research into the theories and uses of social intelligence in robots and virtual agents. Beyond defense and security, the group is slowly expanding towards research areas that involve people’s homes and their normal work environments.

Here, robots will combine the two themes above: they will understand the intents and actions of humans around them, and will be able to collaborate effectively. Applications range from toys through assistance around the house, to mobile robots in industry that work side-by-side with humans.