Insights from AI Expo 2017: Deep Learning in the Enterprise

By Appen. December 06, 2017

I recently attended a keynote session at the AI Expo 2017 called “The Application of Deep Learning within an Enterprise”, a panel featuring speakers across a variety of industries. The executives on the panel highlighted a wide variety of use cases for machine learning in their organizations, beyond what one would expect. For example, Vadim Kutsyy, Distinguished Architect, Data Science from Paypal, discussed how machine learning not only helps Paypal with risk and fraud management, but also ensures that its operations continue to run smoothly, particularly when there are large spikes in traffic such as on Black Friday. With machine learning, Paypal can model how to process the half a pedabyte of data it receives each day. Domenic Venuto, VP and Head of the Consumer Division at The Weather Company, explained how machine learning has helped to boost engagement on its site by 20% by serving users personalized content. The Weather Company has also invested in machine learning for exchange-based media buying, allocating ads and measuring them in real time. The machine learning algorithm is tuned to meet specific KPIs, and dials advertising up and down in real time to meet those metrics. Ali Bouhouch, CTO & VP of Enterprise Architecture at Sephora, discussed how machine learning is used to understand consumer behavior, provide recommendations and optimize transactions. He went further to provide some guidance to other enterprises looking to invest in machine learning, stating that it should not be applied to every problem. He provided this framework:

Look at the problems that your business is trying to solve. Determine which ones are data intensive; those are good targets for machine learning.
It’s not just about taking a manual process and automating it. Look for tasks that have decision-making tied to them and have a cognitive element.
Identify where decisions need to be made in real time and bring in machine learning to enable quick response times.

He wrapped up by stating that the operational challenge with machine learning is how to increase the IQ of the systems more quickly so that the enterprise can make quicker decisions and gain competitive advantage. The various use cases discussed by the panelists demonstrate the wide applicability of machine learning in positioning enterprises for growth. As you prepare to invest in machine learning to drive competitive advantage for your firm, make sure to include high-quality data in your plans. Your investments will pay off more quickly if you start with high-quality data to train your algorithms. Contact us today to learn how we’ve helped companies like Microsoft get more from their machine learning investments.