Using the Dato Toolkits, developers can build software that leverages machine learning technology, combining historical data and real time user interaction to make predictions and decisions. The toolkits allow developers to add recommendations, sentiment analysis, churn prediction and deep learning to create so-called intelligent applications. Dato also announced a partnership with Coursera and the University of Washington to deliver a six-course machine learning curriculum.
"Within the next five years, every successful app will have machine learning as one of its core components" told us Carlos Guestrin, Dato co-founder, CEO and Amazon Professor of Machine Learning at the University of Washington.
"Today, differentiation is going down. Making applications more responsive and adaptive to the individual’s interests will only be possible with machine learning" he added.
When asked about privacy issues, Guestrin notes society has been evolving and according to him, privacy is less of a concern nowadays. Even privacy-aware consumers know where is the value for them of letting machines exploit their data.
Since every service company could potentially be using machine learning within a few years, would historical data eventually be exchanged in a machine-learning ready format or more user-friendly summary so algorithms wouldn’t have to cold-start their learning curve from one service to the next?
And could end-users have a meaningful access to their aggregated profile, say to discover revealing patterns about themselves or uncover recurring tricks being played on them?
Typically, machine learning is only interpreted by computer models that try to provide explanations on the results, explains Guestrin, but in the coming years companies may want to try to make their systems more transparent.
"There is significant research focus on how the results of machine learning for one service could be transferred to another service, but the domains have to be related. What could work for sharing multimedia content would not make sense for your personal medical record".
The CEO doubts that different companies would want to share common profiles, at best companies could re-utilize what they have learned on a profile across different services.
"Machine learning offers an exciting opportunity but we need to make it accessible to more people" he concluded.
What’s in the toolkit?
Recommendation engines can be added to applications to personalize user experiences by mining patterns in purchase and activity history, matching users’ tastes and predicting future purchases or interests. Examples include Pandora, MagazineLuiza and StumbleUpon.
Image search and feature extraction using deep learning can be used to create more intuitive search experiences, automatically tag photos and improve application performance by using highly predictive features extracted from images. Examples include Compology, Beeva and Zillow.
Churn prediction detects which customers are likely to cancel a subscription or service by using advanced analytics to detect specific patterns. It can be added to applications to identify which customers are at risk of "churning". Examples include PayPal and Nuiku.
Sentiment analysis uses natural language processing, text analysis and computational linguistics to identify and extract subjective information in source materials. The technology can be used to discover the tone of a post in support forums or a customer’s opinion of a brand or product in reviews for better targeting. Examples include Nook, Cisco and Moodly.
Download a free 30 days evaluation trial of Dato GraphLab at www.dato.com
Register for the machine learning training through Coursera with prices starting at $79.
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.