Machine Learning - The Future is Here
The term “Machine Learning” might not mean much to you. You might think of a computer program, like AlphaGo playing the board game Go, calculating the multitude of moves and the possible countermoves. However, when you hear the term “Artificial Intelligence” or AI, it’s more likely for you to have visions of Skynet or a dystopian future in which machines rule and that they have created the Matrix to harness the energy of humans. But, the truth about AI and particularly machine learning is far less sinister. It is rather shaping and simplifying the way we live, work, travel, and communicate.
Briefly, Machine Learning (ML) is an application of artificial intelligence that allows systems to learn and improve without being directly programmed. Machine learning has been one of the top tech news topics in recent years and is now being widely applied to businesses. In fact, ML is being used across various disciplines from healthcare to education and it is showing no sign of slowing down. It is hard to ignore the global impact of “AI Washing” in the current business market, and how AI and ML together may change the application-development markets of tomorrow.
Today, AI- and ML-powered systems are drastically changing the way business is done across all industry sectors. Most of you might have heard the story stating how Wal-Mart used machine learning on its customer data to get useful business insights like diapers and beers having a positive correlation. Although this is an urban legend, this is one of the most popular stories in the machine learning domain. So supposedly, Wal-Mart decided to combine the data from its loyalty card system with that from its point of sale systems. The former provided Wal-Mart with demographic data about its customers, the latter told it where, when, and what those customers bought. Once combined, the data was mined extensively and many correlations appeared. Some of these were obvious; such as people who buy gin are also likely to buy tonic. However, one correlation that stood out like a sore thumb. It was found that on Friday evenings young American males who buy diapers also have a predisposition to buy beer. Pertaining to this behavior, the company started keeping the rack of diapers and beers next to each other to increase sales. Nevertheless, the fact is that the most successful businesses today are those which allow certain tasks to be taken over by AI whereby machine learning can acquire more information from and predict consumer behavior.
Another exemplary use of big data and machine learning is done by BMW in the automobile sector to create autonomous cars. BMW is known for designing and building some of the most advanced, high-tech automobiles in the market. So, perhaps it’s no surprise that Big Data-related technology lies at the heart of the company’s business model, from driving decisions and operations across design and engineering, right through to manufacturing, sales, and aftercare. The company believes that the cars of tomorrow will drive themselves, instead of relying on humans being at the wheel. The company has said it plans for its cars to deliver full “level 5” autonomy (the highest level of autonomous cars determined by the U.S. Department of Transport) by 2021. Getting to fully, level 5 autonomous cars relies on computer vision technology, such as that offered by Intel-owned Mobileye. Computer vision means teaching computers and machines to “see” like us humans do, but using cameras instead of eyes, and to interpret what they are seeing in a way that’s similar to the human brain. It’s essentially a very advanced type of image recognition technology. The tech that powers a simple Google image search, where machines are taught to sift (scale-invariant feature transform) through and categorize images, becoming better at it over time as they process more and more images. Only, in this case, instead of sorting through millions of Internet images of cats and dogs, the computer vision in an autonomous vehicle means ingesting all of the input data from the car’s sensors and cameras and analyzing that data on the fly in real-time, quick enough to perform an emergency stop if needed.
However, did you ever think that ML could create music? Well, IBM’s new music algorithm, called “Watson BEAT” has already done it. It helped producer Alex da Kid create the song Not Easy (feat Elle King and Wiz Khalifa). To be clear, IBM’s bot didn’t write the lyrics but rather generated completely new musical scores that Alex used as an inspiration for the tunes. So, let me tell you how it all happened. Watson BEAT first collected music cultural trends and over the last five years, including Nobel Prize speeches, Billboard song lyrics, movie synopses, and more. Then another Watson bot, the Tone Analyzer, examined two million lines of social media content to grasp the emotions around all that data. It then studied recent musical trends, examining the pitch, time, key signatures, and note sequences in songs. Combining that with the emotional and cultural analysis, it was able to create brand new musical scores based on moods like joyful or devastated, or an atmosphere like spooky or cheerful. Based on the analytics, Alex picked up the theme “heartbreak” for his single. After settling on the theme, Alex used Watson BEAT, the cognitive system’s machine learning-driven music generation algorithms, to come up with different musical elements until he found ones that inspired him about how the piece should sound. So, it’s fairer to say the single was inspired by AI, rather than created by it. Nevertheless, this project shows that the technology is there to create a hit record using AI and machine learning. The resulting single, Not Easy reached number four in the iTunes Hot Tracks chart within 48 hours of its release.
Thus, the future culture of work is already upon us as many companies have shifted towards the “community” model of working as the boring tasks are now left to ML and thus decisions are more data-driven. In fact, Microsoft stated in its research published last fall that companies using AI are outperforming by 5% to those which have no AI strategy. As machine learning has provoked worries that our jobs will be replaced by AI, the reality is that Machine Learning is already merely allowing humans to get on with the more interesting facets of their jobs as AI slogs away at the more mundane aspects of operations such as data mining. It’s time for us to embrace machine learning for what it offers us instead of worrying about what it might take away. In the end, we can look at Machine Learning as a time-saving device that allows humans to explore their creative ambitions while ML is in the background crunching numbers and generally taking on the more mundane tasks.
