How is Google using Machine Learning?
Machine Learning will impact everything we do at Google. You can already see the results in Google products you use today. Artificial Intelligence, Machine Learning, Deep Learning, and Big Data are buzzwords we hear every day. What do these words mean? How will Machine Learning make my life better? Let’s take a look.
What is Machine Learning? Isn’t that Artificial Intelligence? Not exactly. Let’s spend a minute defining the terms. AI is a high level term used to describe any approach to make a computer smart. AI started out as a programmed set of rules that could quickly sort through mountains of data to find the desired answer. But, AI rules couldn’t learn or adapt to new data. You could add more “rules” to handle new data, but there was no “learning”. Machine Learning (ML) is a new type of Artificial Intelligence (AI) that lets computers learn without being explicitly programmed. ML is a set of classifiers and algorithms that can teach themselves to grow and adapt when exposed to new data. Machine Learning can learn. This is a BIG deal. Deep Learning (DL) is a particular technique within Machine Learning. It uses an Artificial Neural Network with many layers to learn optimal model parameters. DL creates an algorithm that will automatically decide which features work best to accomplish a task. Finally, Big Data is a term used to describe LOTS of seemingly unrelated data, that upon further analysis could become very useful.
“Machine learning is a core, transformative way by which we’re rethinking how we’re doing everything. We are thoughtfully applying it across all our products, be it search, ads, YouTube, or Play. And we’re in early days, but you will see us — in a systematic way — apply machine learning in all these areas.” Sundar Pichai, Google CEO
We are already seeing Machine Learning embedded in services like Gmail, Search, Maps, and even in what ads we see. Let’s look at some examples in Gmail. Priority Inbox automatically identifies your important incoming messages and separates them out from everything else. It learns over time what is important to you, and what isn’t. Smart Reply is another Inbox feature that suggests up to three responses based on the emails you get. For those emails that only need a quick response, it can take care of the thinking and save time spent typing. This is especially helpful on mobile phones. The responses get better over time as the system learns.
Google Search and Google Maps employ Machine Learning too. When you start typing in the search box it automatically anticipates what you might be looking for and provides suggested search terms. The suggestions could be based on past searches, what is popular now, or where you are at the time.
Google Assistant is a new example of Machine Learning on Android, helping you with everyday tasks. Click on the video at left to see what it can do for you. Google Assistant makes it easy to buy movie tickets while on the go, to find a perfect restaurant for your family to grab a quick bite before the movie starts, and then help you navigate to the theater. Today, Google Assistant is available on the new Google Pixel phone. Older phones have an earlier version called Google Now.
Self Driving Cars are probably the most sophisticated example of Machine Learning in action. If you drive around Mountain View, California you will likely see a Google self-driving car. They have been on campus for several years, and have logged 700,000 miles of accident-free autonomous driving. Watch this video to see what the onboard computers “see” and how they react as they drive the car. It is truly amazing!
Machine Learning will radically transform and improve all kinds of applications. It is already seeping into things like; Security/threat detection, Fraud detection, Recommendation engines for eCommerce, NLP for legal documents, emails, patent searches, Video/voice/speech recognition, Health screening, Cancer detection, Fact checking, Weather or Financial models, and many more.
In an effort to accelerate innovation, Google is making the core Machine Learning framework Tensor Flow available as Open Source. Google Cloud is also providing access to Machine Learning APIs for Natural Language Processing, Speech Recognition, Language Translation, and Vision/Image Recognition.
If you are interested in becoming a Google Partner to take advantage of these and other Google technologies, sign up here.
Next time I will take a look at Virtual Reality and Augmented Reality (AR/VR) and how it will change the world. Click the Follow button to be notified when my next post is available.
Disclosure: Thoughts and opinions expressed here are my own, and do not necessarily reflect those of my employer.
Bio: Don Dodge works on the Google Cloud Technology Partner team. Don is a veteran of five start-ups including Forte Software, AltaVista, Napster, Bowstreet, and Groove Networks. Prior to Google, Don was Director of Business Development for Microsoft’s Emerging Business Team. You can follow Don on Medium, on Twitter @DonDodge or on Facebook