In Machine Learning

Forbes recently noted that Gartner’s 2017 Hype Cycle for Emerging Technologies includes multiple types of artificial intelligence technologies such as machine learning, deep learning, and (new for 2017) deep reinforcement learning. Gartner’s Hype Cycle plots the new technologies that show the most potential for delivering a competitive advantage.

Debuting at the innovation trigger, which is defined as a breakthrough that is driving public interest, reinforcement learning offers a means to find optimal solutions without knowing or explicitly programming in those solutions ahead of time. This can be especially handy for tackling large, complex problems that are beyond the speed and capability of humans to take on alone.

Gartner predicts that deep reinforcement learning will hit mainstream adoption in five to ten years. Interestingly, SpaceTime has been delivering applications based on reinforcement learning to the (typically risk-averse) energy industry for the past four years. These applications have demonstrable business value, improving operations outcomes for a variety of use cases.

For example, a leading wind energy producer uses reinforcement learning to optimize the assignment of its field maintenance crews, taking into account such diverse variables and constraints as wind farm location, proximity and drive time to wind farms, acuity of asset condition, union regulations, skillset availability, and more. Only sophisticated, advanced analytics algorithms like reinforcement learning can effectively process and set optimal crew assignments across hundreds of farms and save thousands of dollars per wind farm per day. No wonder reinforcement learning is rapidly gaining interest!

SpaceTime authored a recent white paper on using reinforcement learning to optimize operations. You can download the white paper here (registration required).

Recent Posts

Leave a Comment

helm_prometheus