Who here remembers taking computer programming in school? Whether you learned programming by punching holes in a never ending series of cards, or by writing simple DOS or other computer language commands, the fact remained that computers needed an incredibly precise set of instructions to accomplish a task.
The more complicated the task, the more complicated your instructions had to be.
Machine learning is inherently different. Rather than telling a computer exactly how to solve a problem, the programmer instead tells it how to go about learning to solve the problem for itself.
Machine learning is really just the very advanced application of statistics to learning to identify patterns in data and then make predictions from those patterns. This website has a gorgeous visualized walkthrough of how machine learning works, if you are interested.
Machine learning started as far back as the 1950s, when computer scientists figured out how to teach a computer to play checkers. From there, as computational power has increased, so has the complexity of the patterns a computer can recognize, and therefore the predictions it can make and problems it can solve.
1. Machines can see.
Because computers are able to look at a large data set and use machine learning algorithms to classify images, it’s relatively easy to write an algorithm that can recognize characteristics in a group of images and categorize them appropriately.
For example, it takes four highly trained medical pathologists to review a breast cancer scan, decide what they’re seeing, and then make a decision about a diagnosis. Now, an algorithm has been written that can detect the cancer more accurately than the best pathologists, freeing the doctors up to make the treatment decisions more quickly and accurately.
The fact that computers can see is also how we get driverless cars. A computer that can recognize the difference between a tree and a pedestrian, a stop and a yield sign, and a road or a field – which is the key to unlocking the promise of the driverless car. And this innovation alone could revolutionize many different business models, from supply chain and delivery to personal transportation.
2. Machines can read.
Google long ago proved the value of a program that can read text. Their search engine algorithm revolutionized Internet search, and continues to do so with every advancement.
But it’s one thing to be able to say whether or not a document contains a certain word or phrase; it’s something else entirely to understand context.
New algorithms are being developed that can determine whether a sentence is positive or negative, context within a document, and more.
In fact, using Google’s street view and its ability to read street numbers, the company was able to map all the addresses in France in just a few hours — a feat that would have taken many talented mapmakers weeks, if not months in the past.
3. Machines can listen.
One of the biggest innovations in recent years is probably in your pocket right now. Siri, Cortana, and Google Now represent a huge leap in machine understanding of human speech.
How many times have you been frustrated trying to get a computer at the other end of a telephone help line to understand you? (I’m sorry, I didn’t catch that… Please repeat your account number…)
Now, virtual personal assistants can recognize a dizzying and ever growing array of commands and respond in kind. More importantly, however, Google and its competitors are moving towards keying their search algorithms to understand natural speech as well, in anticipation of more and more voice search.
In the old days, you would have to type something like, coffee shop + London + a postal code to find a listing of coffee shops in an area. Today, you can type — or speak — a natural sentence like, “Where’s the nearest coffee shop that’s open right now?” and Google understands not only what you mean, but where you are, what time it is, and how to respond.
4. Machines can talk.
Yes, Siri can tell you a knock-knock joke, but that’s not really the kind of talking I’m talking about.
Computer language translations are something of a running joke, and for good reason. There are so many nuances to language — slang, idioms, cultural meaning — that simply running a piece of text through translation software can produce some amusing and ultimately incorrect results.
But new machine learning algorithms are making more accurate, real-time translations possible.
Late last year, Microsoft unveiled real time translations for Skype video conferencing in English and Spanish, with plans to support more than 40 languages.
While the advance in the translation ability is impressive, it’s the combination of listening to the user speak, understanding the words, and translating them all in real time that’s the impressive breakthrough. And because the program is machine learning-based, it will only get better with practice.
5. Machines can write.
While it may take a million monkeys typing to produce the works of Shakespeare, computers are getting a lot better at creative writing.
In one project, a computer was taught to write photo captions describing the pictures. In its first iteration, human readers thought the computer generated description was better than the human generated words one out of four times.
This has broad implications for all kinds of data entry and classification tasks that previously required human intervention. If a computer can recognize something — an image, a document, a file, etc. — and describe it accurately, there could be many uses for such automation.
Another example I have covered before is how during the 2015 Wimbledon tennis championships machine learning algorithms were used to automatically turn match statistics and sensor data collected during each game into automated news stories which read as if they were written by sports journalists.
These skills are beginning to show that computers can now boldly go into realms that were once considered solidly the domain of humans. While the technology still isn’t perfect in many cases, the very concept of machine learning — that machines can continuously and tirelessly improve, they will get better.
How do you foresee machine learning impacting your business or field of work? I’d love to hear your thoughts in the comments below.
Comments will be approved before showing up.