Every Monday morning, 75 million Spotify users logon to their devices and play a "Discover Weekly" playlist that's been curated to their musical preferences. Unbeknownst to many, the custom playlist they're enjoying was curated by a machine learning algorithm that has learned their unique musical preferences based on previous interactions with songs, musicians, and playlists.
Machine learning, and a more advanced technology called deep learning, are types of artificial intelligence that allow a computer to learn information based on the data it is given. To borrow the words of Drew Breunig, "In a nutshell, deep learning is human recognition at computer scale."
Essentially, the more information the computer is given, the better it can learn -- and in the case of platforms like Spotify or Netflix, the more interaction you have with the program, the better it can recommend music, movies, or TV shows that you'll like.
It's no coincidence that Facebook suggests which friends to tag in photos, or Amazon recommends the products you'll like. The recommendations are based on machine learning algorithms that have learned your patterns and preferences in order to tailor content specifically to you.
Given the level of investment in machine learning by companies like Google, Facebook, Amazon, and Microsoft, it's no surprise that these technologies will redefine our society, and more importantly, the way that businesses operate. As John Giannandrea, Google's Head of Machine Learning, put it "we're in an AI spring."
Bots For Business
Bots are not an entirely new concept: when you call your bank or phone service provider you're asked via automated message to specify your reason for calling before speaking with a representative -- this is an example of natural language processing. Applications like Siri, Cortana, and Alexa help keep our personal lives organized by making appointments, setting alarms, and sending messages when we're not able to, all via verbal commands.
Soon, chatting with a bot will be like speaking with another person...
To that end, Viv, an AI virtual system, was created to go further than other language processing applications through its ability to respond to follow-up questions with total accuracy. That means you can do more than just check the weather -- you'll be able to ask Viv complex questions and get real responses, like "Will it be warmer than 25-degrees near the at the Rogers Centre after 7 p.m. the day after tomorrow?"
By leveraging natural language processing with machine learning algorithms, businesses can manage their entire customer service process using bots. Earlier this year, Sephora launched a chatbot on Kik aimed at younger millennials to get to know the user and serve relevant content in order to lead the customer towards making a purchase, without requiring them to leave the app. As noted by Chris Messina, "beyond human-to-human dialogue, it's bots that enable this to occur at scale."
In fact, just last week Microsoft added five new chatbots available to Skype users for Skyscanner, Hipmunk, and Stubhub. The programmable messaging bot will give users notifications about the information they're looking for on their connected devices, or social media platforms.
Soon, chatting with a bot will be like speaking with another person -- at least that's the intention behind Google and Facebook's advancement in deep text. Their work aims to improve a bot's neural network speech patterns, understanding sentence grammar, and an overall understanding the meaning of a conversation.
Learning Customer Behaviour
Dozens of banks across Europe already use machine learning applications, and in many cases, have experienced double-digit increases in sales, cash collection, and significant decline in customer churn. These algorithms have also enabled the creation of loan default prediction models, like the one we developed to predict customer loan default with 99.1 per cent accuracy for a credit risk insurance company.
Machine learning provides insight into a business' customer base, predicting customer interactions, if or when a customer is likely to switch to a competitor, and when they'll be interested in making a new purchase or interaction with a company.
Ramping Up Cyber Security
With the amount of data moving around all the time, it can be difficult for a business to protect its data. By applying the right machine learning algorithms, a business can track cyber risks and deploy alerts and solutions in real-time to mitigate downstream effects of a data breach, as well as identify suspicious transactions with payment providers.
Defining A New Era
As data becomes more and more accessible, machine learning algorithms will continue to create efficiencies and improve our everyday lives -- way beyond better song or movie recommendations.
Within the next year, you may start to notice a more personalized touch from your customer service "representative," a stronger relationship with your teaching assistant, and better investment advice from your financial analyst. So, when you encounter these scenarios and many new ones, remember that you may be dealing with artificial intelligence -- powered by your own data.
"It's not magic, it's just a tool. But it's a really important tool," -- Greg Corrado, senior research scientist, Google
Follow HuffPost Canada Blogs on Facebook
ALSO ON HUFFPOST:
Follow David Suydam on Twitter: www.twitter.com/DSuydam