KDnuggets Home » News » 2019 » Sep » Meetings » Will Machine Learning End Retail? Data Science Seattle Oct 17, 2019 ( 19:n37 )

Will Machine Learning End Retail? Data Science Seattle Oct 17, 2019


In advance of the Data Science Salon taking place in Seattle on Oct 17, we asked our speakers to shed some light on how Artificial Intelligence and Machine Learning are impacting one of America’s most disruptive industries. Read for more insight, and then register with KDnuggets exclusive link for 20% off tickets.



Sponsored Post.

Figure

Some say the rise of eCommerce portends the fall of Retail. But both Retail and eCommerce increasingly rely on progressively sophisticated Machine Learning solutions to get better business insights and drive demand.

In advance of the Data Science Salon taking place in Seattle on October 17, we asked our speakers to shed some light on how Artificial Intelligence and Machine Learning are impacting one of America’s most disruptive industries – retail and ecommerce. There’s no doubt that the landscape is changing quickly, and the next big ideas are being driven by most robust ML deployments.

New ML Efficiencies 

We’re definitely seeing the needle move on the business analytics side according to Dhivya Rajprasad, Data Scientist at Levi Strauss & Co. “ML is slowly moving from primarily in the area of technology to retail from inventory forecasts to assortment planning to empower and make business decisions which rely on intuition to becoming more data driven.” But customers are seeing the benefit too. “Our entire business exists to create tools to address big inefficiencies that currently exist in shopping,” said Rhonda Textor, Head of Data Science at True Fit. “Online shopping provides large catalogs of items for shoppers, but that comes with increased difficulty in finding what you are looking for.  Also, the lack of ability to interact with clothing means that shoppers are more likely to return items that they buy online. We use ML to recommend items we think shoppers will like, and we use ML to help shoppers get the right size, which decreases the need to return items.”

Privacy & Ethics

It’s also clear that when it comes to privacy, retailers have taken notice. “Models are affecting lives, and we need to understand and consider the impact,” said Textor. Rajprasad agrees, “Privacy and ethics are and should be the cornerstone of data science vision for a company. One challenge we face is ensuring the highest level of privacy by not encoding real time PII data which will be still be connected but completely anonymized to ensure a degree of personalization without compromising the privacy of the person.” Textor also agrees with this focus on anonymization. “In order to receive personalized recommendations, our users provide very personal information such as height and weight.  We have a responsibility to protect that data, to ensure that the raw data is not exposed with any personally identifying information, and to provide value to the user in the form of good recommendations. We are also working to provide experiences for shoppers who elect to remain anonymous by providing general information about products that do not require information specific to shoppers.”

But there’s also ethical quandaries embedded in the model creation process. “[Another] common challenge is human biases when building a model which could probably serve very well for eg : an American market but will fail in Asian markets because the model was not trained considering all kinds of population in mind,” said Rajprasad. 

The Next Step for the Community

It’s never easy to predict the future, and our speakers had divergent opinions on its direction. “This is the age of data democracy and I strongly believe in open sourcing (essentially removing the middle men) all our work to enable the world to be a better place using data,” said Rajprasad. “There is definitely going to be a huge convergence and possibly a central council in every company to ensure data is being used in the right way and by the right people.”

Textor disagrees, “I think it's too early to think about a convergence of tools and methods.  There have been many exciting advances in recommendation and computer vision technologies that are powering new experiences such as visual search in shopping.  It's too early to know where those experiences are succeeding and where they fall short. As the academic and research communities continue to rapidly advance the basic science in this area, we will continue to try out new tools.”

Where can I learn more?

Make sure to catch all of our speakers at the Data Science Salon Seattle on Thursday, October 17. Tickets are almost sold out, so pick yours up today! Follow this link for 20% off tickets, exclusive to KDnuggets readers.


Sign Up

By subscribing you accept KDnuggets Privacy Policy