There are many in this world that would swear blind artificial intelligence (AI) will never happen. However, there are also a lot of people that would say it is already here and when you take a look at some of the world’s latest inventions, you might well be hard pressed to say that AI isn’t already in our midst.
One such invention is called DuLight and is the latest protype from Chinese tech giant, Baidu. The device is tiny but packs a big punch for it can enable the blind to see!
How does DuLight work?
Well, the device has a tiny camera built into it, which can capture the image of whatever is positioned in front of you, whether that image is a human face, food or a computer. The device then transmits these images to a smartphone app, which, in turn, analyses the images and concludes what they depict. The app then generates an audio description of the image through to an earpiece being worn by you (a blind/sight impaired person).
The prototype uses a technology known as deep learning. It works via the process of a huge number of digital images being fed through Baidu’s online neural network. Eventually, if enough are fed through, the neural networks learn how to identify the images. The images can be of simple things such as clouds, right through to words as well as human faces.
Of course, DuLight is still at prototype stage, so it is not working the way it eventually will do. Nonetheless, it clearly shows how machines are becoming ever more advanced and how we are becoming ever more used to the idea that eventually they will be able to view and actually understand their own surroundings.
Deep learning technology
Baidu is not, of course, the first to use this technology. Other companies such as Google, Microsoft and Facebook also use neural networks in order to help you search for and identify a range of things, from people to places. The technology helps your smartphone assistant answer your questions and respond to your commands. Many will also be familiar with Facebook’s face recognition feature, as well as Skype’s instant language translation feature. Both of which are based on deep learning technology.
Neural nets are, more and more, becoming used to the many different methods of perception that humans use on a daily basis, even the way we speak. For instance Google has recently created a chatbot that can hold a conversation with a human simply from having been fed various movie dialogues into its memory. The chatbot learns how to respond to questions posed to it from the movie dialogues it has been fed.
There is no doubt that deep learning technology will advance a lot in the year ahead, especially with tech giants, such as Google and Facebook on side to its continuing development. For instance Google recently made the move to open source the software engine that drives its own deep learning services, and Facebook has now open sourced its designs for its own hardware server that drives its deep learning work.
Indeed, it seem that Google started the ball rolling for not long after it made its own announcement, Facebook made its announcement, and a day later, OpenAI, a $1 billion non-profit was announced by a group, which is led by Elon Musk (Tesla Motors founder) and Sam Altman (President of Y Combinator).
They announced that the non-profit would be open to sharing all of the research and technology they have on AI with the rest of the world. This news is, of course, big in its own right but more so because Ilya Sutskever, a former AI researcher at Google, is overseeing it. The non-profit has also been pitched as something that will help protect the world from the dangers of AI, i.e. of AI coming too powerful and beyond the control of humans.
There will always be those who worry about the dangers of AI, and there is no doubt that these dangers should be considered carefully. However, although AI has come on leaps and bounds over the past few years, we are still many years away from it being powerful enough to be classed as super-human AI. For now, we should be focused on the extreme good that the technology can do, such as with DuLight.
So for those who don’t believe AI is real yet, you only have to look at the features Facebook, Microsoft and Google are already experimenting with to see it is real and already! There may still be some way to go before it reaches its peak but it is here and most people will be using it on a daily basis, perhaps without even realising exactly what it is they are using.