Scientists at the Institute of Neuroinformatics at the University of Zurich in Switzerland, are working on training a robot to hunt its prey. That s right, this is the cue to panic! (Think, iRobot and Blade Runner). The first thing that comes to my mind when I think of robots learning to hunt, is me running from angry robots that are trying to erase humans from the planet. Also Read - Realme announces 'D' under its TechLife division; will focus on smart home devicesAlso Read - Sony leading image sensor market, Samsung still trails behind
However, the scientists say that the hunting lesson for robots is not as freaky as I make it sounds. When scientists train a robot to identify its prey and predator, it s essentially supposed to go up to levels of understanding a parent and child relationship. Also Read - Instagram uses AI to automatically hide offensive comments
For example, if you think of an automated shopping cart in the future that follows the customer across the store, it would be a robot trained to understand its prey, which is this case would be the customer, so that it knows to follow it till instructed. On the other hand, a predator would probably be the shelves and boxes and people, that the robotic cart must not run into.
The image below shows the raw data representing what a predator robot sees.
Interestingly, these robots are modeled after a member of the animal kingdom. The robot also features a silicon eye , just like a human eye. This silicon eyes allows the robot to detect and process the movement of its prey in real time, instead of a frame by frame series in a regular camera.
This data that the robot receives is then processed using neural network, which is essentially teaching the artificial intelligence (AI) device to learn how to react from a particular input, so that if a similar input is given to it in future, it knows what to do.
This is not the first time that scientists are working on a robot or AI to learn more human-like habits. Recently, an AI, which named itself Benjamin, wrote a screenplay for a sci-fi short film, called Sunspring. It was created by technologist Ross Goodwin and was authored by a recurrent neural network called long short-term memory (LSTM). For the script, the AI was fed data of sci-fi screenplays mostly from the 1980s and 90s, and 30,000 pop songs after which it also composed a song that is featured in the movie.