4 Key Trends for Artificial Intelligence in 2017


Last year was huge for advancements in artificial intelligence and machine learning. 2017 could be the year of artificial intelligence and machine learning. Here are five key trends to look out for;

Chinese AI boom
China is poised to become a major player in AI as the country’s huge investments in innovation bear fruit. China no longer needs to rely on copying western tech and is stepping out on her own. Baidu, the Chinese Google, has had its own AI lab for some time and is making big progress in voice recognition and natural language processing. Tencent which owns the successful mobile messaging app WeChat, opened an AI lab last year. Didi, which bought out Uber’s Chinese is currently establishing its own lab and working on its own driverless cars. AI-focused startups are hot buys for Chinese investors, and the Chinese government is sinking in $15bn into AI over the next 24 months.

Language is a hot area. Researchers believe that the strong progress we have made in voice and image recognition, and other areas, will help machines to make significant progress in language. There’s still some way to go, given the complexity and subtlety of language but this is a major focus given the tremendous boost to usefulness that improved language abilities would give machines. Significant progress is already being made, and we can expect a whole lot more in 2017.

Positive reinforcement
Reinforcement learning is based on how animals learn as a result of positive or negative outcomes. This allows machines to learn without instruction or a master-plan. This is not a new idea, but now we are able to combine it with deep neural networks machines have the ability to tackle complex problems. Machines are now able to employ a mix of relentless experimentation with an analysis of previous problem solving to achieve success in a vast number of situations. Scientists are now pondering the extent to which reinforcement learning can be applied in the real world. The development of simulations to allow machines to learn in a safe environment could boost progress significantly. Could positive reinforcement learning be at the heart of 2017’s next big thing?

Generative Adversarial Networks
Generative adversarial networks, or GANs, are systems consisting of one network that generates new data that it learns from a training set, and another that can tell the difference between real and artificial data. These networks can produce very realistic synthetic data by working together. Advances in this field could be used to create more realistic video gaming visuals, improve video and improve computer generated design.