What will be the next thing to revolutionize data science in 2019? originally appeared on Quora: the place to gain and share knowledge, empowering people to learn from others and better understand the world.
Answer by Håkon Hapnes Strand, Data Scientist, on Quora:
Reinforcement learning will be the next big thing in data science in 2019.
While RL has been around for a long time in academia, it has hardly seen any industry adoption at all.
Why? Partly because there have been plenty of low-hanging fruits to pick in predictive analytics, but mostly because of the barriers to implementation, knowledge and available tools.
The potential value in using RL in proactive analytics and AI is enormous, but it also demands a greater skill set to master. Not only does it involve more complex algorithms and less mature tools, but it also requires accurate simulations of real-world environments. An increasing number of people in the industry know about the unlimited potential of RL, but very few have been willing to make actual investments. It’s mostly considered too risky.
All of that is about to change.
The widely published successes of DeepMind’s AlphaZero and OpenAI’s Dota bot have opened even more eyes to the almost magical capabilities of deep reinforcement learning agents. This buzz is finally beginning to materialize into something real. There are whispers of companies investing in R&D and letting their data scientists spend time on RL proof of concepts.
Better tools are also right around the corner. Microsoft recently acquired the startup Bonsai AI that has built a generic DRL platform, and you know when Microsoft gets involved it means business.
There’s also a generation of data scientists who have been around for some years now and are getting tired of the old “use inputs X to predict target Y” problems. RL offers something new and exciting. DRL is currently our best hope in the push towards AGI, and that makes a lot of people excited too.
Pretending like RL will go completely mainstream in 2019 is naive, but it will take a big step forward. The most likely early adopters are manufacturing, robotics, transportation, and e-commerce. Once it does go mainstream, it will revolutionize everything.
The Nerd's Stash 