You don't need to look very far to see examples of how machine- and deep-learning could enrich our lives. You might be surprised, given how often we hear about deep-learning nowadays, that most companies have only been utilizing the technology to a notable degree for the past couple of years - including Amazon. Despite that, we've seen some amazing things accomplished in a relatively short time.
With Amazon specifically, we've seen the company sell the fruits of its deep-learning labor to those who need reliable predictive services, use it to improve its product reviews sections, and of course, improve the usefulness of its Alexa AI, found in products like Amazon Echo. Today, Amazon is giving us another reason to warm up to the idea of AI and what can be done with it.
Amazon's Echo takes advantage of the company's deep-learning prowess
In this unveil, we looking at an example similar to one from Microsoft in 2015: age-guessing. This latest development is tied into Amazon's new Rekognition AI service, built by Amazon's Computer Vision team. As the name suggests, it is used to help computers recognize real-world objects and places, as well as faces.
Amazon's Jeff Barr wrote up a great post on how Rekognition works, and went as far as using his own photos from decades past, to prove how well it works. With his baby photo seen above, Rekognition was able to identify that there is a face in the picture, with 99.7% assurance, and that the face is male, with a much less confident 62.2%. It further determined that the boy was between 0 - 2 years of age, and that he was smiling - both spot-on results.
It's obvious that determining the sex of babies is quite difficult, and let's be honest: it's sometimes difficult for actual humans to determine, as well. But with adults, Rekognition's chance of a successful guess dramatically increases, with other sample pictures of Barr presented, increasing the confidence level to 99.7%. It might be easier to guess that a baby is between 0 - 2 years, although those ranges vary quite dramatically as a person gets older, with one particular photo placing Barr between 35 and 52 years old.
Amazon's Rekognition toolset is exclusive to developers
In some cases, Rekognition could prove to be a better guesser than a real human, while in others, the advantage could flip-flop. However, what's important to stress is that this is only what Amazon's service is capable of right now. We're still early in this deep-learning game, so there's no doubt that these guesses will improve over time, as Amazon continues to churn through its never-ending supply of data, feeding the machine.
If there's one downside to this awesome technology, it's that Amazon has locked its age-guesser tool behind its AWS developer service. This requires a credit card to sign up, and while no cost will be charged simply from signing up, the requirement will undoubtedly prevent most mainstream users from testing it out. What we can hope for is that Amazon sees the excitement around the tool and releases its own public demo, or that a developer decides to build on top of the functionality and release their own take.
Even without being able to use the tool, though, it's easy to appreciate what's become possible in so little time through deep-learning.