Machine Learning technologies are critical to the design, manufacture, management and improvement of modern industrial assets. At GE Research we are infusing advanced Machine Learning algorithms into all aspects of GE's industrial portfolio to enable superior product design and more intelligent asset management.
We're enhancing GE's assets and services with adaptive algorithms that continuously learn from real and virtual experiences to improve operational efficiency in dynamic environments. We're able to turn large amounts of structured and unstructured sensor data from GE's test platforms, installed assets and data recorded during maintenance processes into actionable insights. These insights ultimately improve operational efficiency, eliminate unplanned down-time and reduce maintenance costs.
Working with a wealth of data across multiple modalities and deep industry domain knowledge, GE Research develops novel unsupervised, semi-supervised and supervised, as well as reinforcement learning technologies to address critical challenges for our customers that deliver differentiated solutions. We focus on critical areas such as fault diagnostics, prognostics, process automation and prescriptive analytics that can quantify uncertainty and assess risk, to provide safe and optimal actions or policies to improve business outcomes.
GE Research is also creating a new field called “humble” AI, which can understand its own region of trust and improve its competence over time via continuous learning.