2016 can surely be called the year of AI, yet we imagine it to develop and expand even more in 2017. If you are interested, take a look at the examples below.

Reinforcement learning

Have you ever paid attention to the way animals tend to learn certain behaviours when they are motivated? The ‘live it, learn it’ saying sums this approach accurately.

Apparently, computers can learn behaviours and identify objects. They learn by remembering the whole process of certain tasks, that is the outcome and the actions that preceded it. This allows the machines to learn through trial and error without being given any instructions at all.

The process is an example of machine learning connected with deep learning.

Hopefully, the reinforcement technique will be available to use real-life scenarios since the data accumulated will be useful now, as well as in years ahead of us. Not only can we expect computers and applications to be better suited to our needs, but they will also help shape our lives on a more tangible level.

Generative Adversarial Networks (GAN)

Back in 2014, Ian Goodfellow came up with new unsupervised machine learning technique – Generative Adversarial Networks.

For those who forgot, the unsupervised techniques are based on teaching the machine by giving it large volumes of data to finally make it able to identify things based on different aspects.

The tests incorporate two networks, one of them creates analysis based on learnt data and the other attempts to distinguish between real and false data. Together they can come up with valid results.

It is still being tested, however, it can definitely be used in producing photorealistic images in games to visualise objects.

Artificial General Intelligence (AGI)

Artificial Intelligence (AI) isn’t anything new and has been constantly growing, with its momentum being right now. Nevertheless, it developed into something similar yet slightly different – precisely the Artificial General Intelligence (AGI). It is the ability of machines to perform tasks on the same level of intelligence as a human being.

The AGI is often correlated with robotics and is supposed to create more human-like robots that will be able to perform exactly the same actions as flesh and blood humans. Examples are endless – passing exams or preparing coffee made to order.

It is still the beginning of this emerging technology but we may be able to see its slight growth during 2017, unfortunately its full potential may not be perceivable until 2050.

2017 may be extremely fruitful for Artificial Intelligence and the new developments may change the way we think about the world. One thing is certain – it is still just the beginning of its flourish.