Key observations relevant to geoscience use of machine learning are as follows:

Data Prep - for front end data preparation and processing, Azure ML Studio stands out. It looks and feels like modular signal processing design. So anyone who has seismic data processing background will find a natural fit. Coupled with Excel user interface (next), it paves the way to max utilization.

User Interface for Machine Learning Input Parameters - Azure ML Studio ranks at the top because: a) option to use Excel as front end to its web services, b) in comfortable and familiar interface for engineers, financial folks and geoscientists, without writing any browser extensible markup language (XML) codes, hence c) bodes well for rapid experimentation and deployment.

Recall in June post, “the rate of innovation: external open-source exceed that of internal org-specific, we are getting there fast.”

Acceleration in the marketplace - while Azure has been the platform of choice for hybrid computing, a mix of on-premise and cloud, that flexibility is now available on all three platforms (GCP Q3 and AWS joins in this week). Beyond that, all have Machine Learning Apps Marketplace.

The dream of having plug-n-play inter-operable applications is fast becoming reality. Expect machine learning apps become utility, with increasing choices and flexibility in 2019.