Ben Hamm, an employee of Amazon used machine-learning software to create a system that, he says, will recognise when a cat is holding a dead animal in its mouth. In the event that it is, the system will lock the cat out for 15 minutes.
According to theBBC, Hamm claims the device, which he unveiled in Seattle in June, has locked his cat out successfully six out of seven times. Although he also states that it is not 100% fool proof.
Hamm’s AI software forms part of a swathe of work by amateur scientists applying machine-learning to various aspects of everyday life.
Hamm needed to feed into his system thousands of pictures of his cat Metric walking to and from the door, and carrying birds and mice. Once the computer was fed these images it could then, theoretically, learn to distinguish between a victorious Metric and one that arrives at the door empty-handed.
According to the BBC Hamm used two different types of software:
• DeepLens - a video camera specifically designed to be used in machine-learning experiments.
• Sagemaker - a service that allows customers to either buy third-party algorithms or to build their own, then train and tune them with their own data, and finally put them to use.
The main problem encountered by Hamm was the number of images required to “teach” the system to recognise the cat. In more complex facial recognition systems millions of images are required of a single subject in order to yield an accurate result.
Some police authorities in the US have come under fire for their misguided use of facial recognition. However, facial recognition software is transforming security systems worldwide, especially those of commercial airlines.