Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Robots need to understand motive like humans, new report says

Machines should understand motive the way humans do, and not just perform tasks blindly, a new study claimed.

The article by the National Centre for Nuclear Robotics (NCNR), based at the University of Birmingham, is based on research by lead author Dr Valerio Ortenzi.

In the paper, which was published in Nature Machine Intelligence last week, Ortenzi explores the ‘motive’ issue by analysing robots’ ability to grasp objects.

According to the researcher, the majority of factory-based machines are ‘dumb’, in the sense that they blindly pick up familiar objects that appear in pre-determined places at the right moment. 

In order for a machine to pick up randomly-presented unfamiliar objects, technologies such as vision systems and advanced AI must be deployed. The machine can then see the target and determine its physical properties, potentially adjusting its gripping sensors not to inadvertently crush an object it has been told to pick up.

However, Ortenzi argues in the paper, even in the scenario of these technologically advanced robots, the ‘motive’ issue persists. In other words, a traditionally ‘successful’ grasp for a robot might actually result in a real-world failure, since the machine does not take into account what the goal behind picking up that object is.

“Imagine asking a robot to pass you a screwdriver in a workshop,” the study reads. “Based on current conventions the best way for a robot to pick up the tool is by the handle.

“Unfortunately, that could mean that a hugely powerful machine then thrusts a potentially lethal blade towards you, at speed. Instead, the robot needs to know what the end goal is, i.e. to pass the screwdriver safely to its human colleague, in order to rethink its actions.”

The report examines various other examples of this issues, including a robot correctly handling a glass of water to give to a resident in a care home and an another one in a factory correctly picking up an object for delivery but by doing so obscuring a crucial barcode, thus making the delivery impossible.

“What is obvious to humans has to be programmed into a machine and this requires a profoundly different approach”, the researchers said. “The traditional metrics used by researchers, over the past twenty years, to assess robotic manipulation, are not sufficient. In the most practical sense, robots need a new philosophy to get a grip.”

To purchase and read the report in full, you can check out Nature Machine Intelligence’s website here.

Image via PXhere and Max Pixel.


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter