Finding the killer cloud-based machine learning applications

August 16, 2018

A blog post by David Linthicum, managing director, chief cloud strategy officer, Deloitte Consulting LLP

But, now that we have it, what do we do with it? There is an ongoing push to find just the right use cases for machine learning and the cloud. Here are a few concept suggestions you might want to explore:

Risk analytics

Most businesses want to understand the risk of doing things, such as investments, pushing out new products or services, or perhaps the risk of taking on specific customers. These are calculations that people deal with all of the time in their heads, but they are often wrong since all things can never be taken into consideration.

The ability to deal with risk via analytics means that we consider all of the relevant data that's available. In the world of finance, this could be the history of a stock we're about to trade, or economic data, or both.

The role of machine learning is to take all of the data into account, no matter how diverse, as we make a risk assessment. Machine learning also has the ability to learn as it processes, not only from the data it has access to, but from experience algorithms that are built into a constantly improving knowledge-base, much like human intelligence.

Marketing

The ability to find a market, as well as create demand in that market, has been more art than science for the last 100 years or so. Machine learning brings a new dimension with systems that can take historical data into account and build experiences as to what works within a market, and what does not.

Machine learning actually deals with data complexity for you, if properly set up. It can deal with the massive amounts of data, and how that data can be culled and sorted to reach logical conclusions that are equivalent to thousands of experts working together to make a single set of recommendations.

Case in point, let's say you're looking to potentially bring a product to market. You need to consider history, which includes the successes and failures of similar products brought to market. Also, relevant data such as economic factors, changing demographics, sales patterns by region, etc.

This is likely too much data to be compiled by a single human or sets of humans. Considering the amount of data and numbers of patterns that need to be identified and leveraged, this is where machine learning systems really shine.

Picking a path

Cloud-based machine learning is another tool in the shed to help improve your business. Like any tool, it can be misapplied and you will not obtain its full potential.

The trick is to pick the right applications for this technology. This typically means automating analytics where, as the amount of data increases, the analytics actually work better.That does not happen with humans. In fact, it's the reverse.