Imagine a robot that could clean your house, work out immediate repairs, wash the linen, buy the barbeque sauce for the party that you are throwing, and power up your home theatre system for the evening, all without any prior instructions. This robot understands your habits, preferences, and learns from its surroundings and functions accordingly – very similar to a human housekeeper.
The artificial intelligence (AI) that we associate most robots with at present requires detailed programming. However, while this works for linear execution of written programs, these robots are unable to understand additional inputs without prior programming and are reaching their performance limits.
This is where neuro-inspired hardware, or neuromorphic chips, can have enormous impact and enhance AI. In a fast-paced world that involves juggling multiple responsibilities, robots with neuromorphic chips or microprocessors can grasp information like humans, improving the overall efficacy of robotics immensely.
Neuromorphic chips not only mirror brain-biology and functioning, but also aim to mimic brain-behaviour. Fashioned out of silicon, in a manner similar to the human central nervous system, these chips enable robots, smartphones and other vision systems to acquire cognitive abilities. For instance, drones with neuromorphic chips will not need maps to be pre-programmed with precise route information to avoid unfortunate accidents with buildings or birds. Fitted with a sensor that enables 180 degree vision, these drones will be able to process images of their surroundings, recognize obstructions, and avoid collisions—instantly improving their efficiency.
Neurons in the human body modify their connections based on the stimuli they receive, allowing the body to respond appropriately enabling learnt behaviour. The neuromorphic chip contains neural networks that emulate this behaviour using sensory units that provide external information. A robot embedded with such a powerful chip can ‘watch’ a cardiac surgeon dealing with a complication in a routine surgical procedure and replicate the actions precisely by learning from its visual feed, and is also armed with the knowledge of dealing with any other issues that could arise.
With continuous improvement in the performance of neuromorphic chips, you could even train your phone to take a picture of your daughter, when she is on stage during a school play. By processing images, sounds and other sensory information, your phone will be able to recognize your child’s face accurately and take a picture for you.
In a manufacturing set-up, these high capability robots could greatly augment production and speed. Neuro-inspired technology enables a robotic device to perceive things as humans would, allowing it to not only function with minimal instruction, but also anticipate future requirements and predict patterns in complex data. Equipped with these chips, devices can become our cognitive partners and can learn to adapt and respond according to our surroundings and habits.
What are your thoughts on the cognitive capabilities of robots? Please share your comments in the section below.