Machine Learning and Artificial Intelligence in 2016

2016 was an interesting year. AI winter is over, but this time AI is almost a synonym for deep learning. Major technology companies (Google, Microsoft, Facebook, Amazon and Apple) announced new products and services built using machine learning. DeepMind AlphaGo beat the world champion in Go. Salesforce bought MetaMind to build a deep learning lab. Apple promised to open up its deep learning research.

Stanford started One Hundred Year Study of Artificial Intelligence.

NIPS 2016 (Neural Information Processing Systems) conference had 2500+ papers submitted and 5000+ people in attendance. A lot of exciting research was published last year. Overview of NIPS 2016: day 0&1, day 2, day 3.

Bay Area Deep Learning School was a good way to learn about state of the art in different areas of deep learning research: computer vision, NLP, unsupervised learning, tips and tricks for applying deep learning. It was a two-day event with great speakers, including Andrew Ng, Yoshua Bengio, Andrej Karpathy, Richard Socher, Russ Salakhutdinov and others. The video is available on YouTube: day 1, day 2.

Chatbots are on the rise. Facebook rolled out bot API for Messenger. Microsoft made it possible to build bots for Skype. Google bought Facebook bought Microsoft built Bot Framework and Microsoft launched and shut down Later they quietly rolled out the next version of the chatbot, called Zo.

Bot builder community is large, Bots group in Facebook has more than 20,000 members, but there are very few really good use cases. Bot builders are actively exploring the space trying to figure out what use cases are good for bots, how to design dialogs, how to make bots smarter. Smartness is still an unsolved problem, interfaces of the majority of the bots are based on buttons rather than natural language understanding.

Self-driving cars are a hot topic. 2016 news: Tesla announced the self-driving car plan, GM, BMW and Ford are building self-driving cars, Apple pulls back, and Uber launches and then cancels self-driving cars in San Francisco due to troubles with regulation. Self-driving car technology is rapidly improving, and regulation will be the main obstacle for a wider roll-out of the technology.

AI will replace many jobs, probably starting from truck drivers, but also knowledge jobs. It is already happening, and even hedge fund managers can be replaced. Some people are scared of this transition, some are excited, and some try to think how to guide the transition to get to the better future. AI is coming no matter what we think about it, so let’s learn how to make it useful.


Am I missing any interesting areas of research? Please share in comments!