Using Artificial Intelligence to Grow Your Business

In my last post, I discussed the importance of big data and how companies are capturing more data than ever to make better and faster decisions.

How, though, are companies putting this data to work?

In this post, I will discuss how big data is empowering the artificial intelligence revolution, the differences between artificial intelligence, machine learning, and deep learning, and how your company can take advantage of these technologies.

Why Now?

Artificial intelligence is a very broad concept and simply refers to intelligent machines.

The term dates back to the 1956 Dartmouth Conference, when a bunch of computer scientists got together to explore the possibility that “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” For decades, AI was more of a science fiction concept (think “2001: A Space Odyssey” or “The Terminator”) than anything else.

However, in recent years, the amount of data that has been captured coupled with advances in computational power has sparked a new wave of AI innovations. And this time, it’s not science fiction.

What Can Artificial Intelligence Do?

There are three types of AI:

1. Artificial Narrow Intelligence: ANI, or weak AI, is AI that specializes in one specific task, often outperforming humans in this task. ANI is the computer that can beat anyone in the world at chess, every time. Ask it to do anything else, such as recommend a movie, and it doesn't have a clue what you’re talking about.
2. Artificial General Intelligence: AGI, strong AI, or human-level AI, refers to a computer than can perform any intellectual task that a human can. Accomplishing AGI would mean that a computer could reason, plan, solve problems, think abstractly, and comprehend complex ideas. We are not yet here.
3. Artificial Superintelligence: Nick Bostrom describes ASI as a computer “that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom, and social skills.” Elon Musk, Stephen Hawking, Bill Gates, and many others have warned of the perils associated with ASI. Thankfully, we are nowhere close to being here.

ANI is everywhere and you probably interact with it several times throughout your day -- think Siri, Spotify music recommendations, and the text that your bank sent you when it thought that fraud had occurred. I won’t dive too deeply into the descriptions of AI but for my favorite overview, take a look at Tim Urban’s 2 part series on AI.

What is Machine Learning?

Machine learning refers to a computer’s ability to learn and therefore become smarter and smarter. Companies, like Google, are enabling machines to learn how to learn so they can make inferences based on their observations.

How does this work?

Well, think about how humans learn. We learn from examples of things and repeated practice -- this is not much different than how machines learn. Machines learn by finding patterns in large amounts of input data. The goal of this is to create a predictable model that can generalize a particular set of data in order to solve a problem.

And Deep Learning?

Deep learning is a specific kind of machine learning that takes advantage of neural networks. It basically builds layers upon layers of information as part of the learning process.

Let’s take a simple case of a computer learning whether a picture contains a cat or a dog. The first layer of the neural network might ask if the picture contains a splotch of brown. As we progress up through the neural network, the layers get more complicated, asking questions like “does the splotch of color have whiskers?”

One of the neat aspects of deep learning is that the programmers don't have to program into the algorithm that cats have whiskers. Rather, the machine learns that in most photos with cats, these whisker type things occur.

So, Cats and Dogs...What's the Big Deal?

There is a computer vision contest every year. The winner of this contest is the team that can identify images taken from thousands of different categories with the lowest error rate. In 2011, the winning team had a 26% error rate; humans had a 5% error rate. In the 2016 contest, the winning team had a 3% error rate, surpassing the rate of humans.

Yes, describing a picture as a cat or a dog might not be the best example of how deep learning can better the future of humanity and Facebook’s ability to tag you automatically because they have learned what your face looks like is not going to make or break your day. But, how about when deep learning is applied to healthcare?

Let’s take diabetic retinopathy as an example.

Diabetic retinopathy is the fastest growing cause of blindness and there are more than 400 million people at risk worldwide. The predicament about diabetic retinopathy is that it is tough to screen for it. To screen for the condition, doctors simply look into someone’s eyes and determine whether or not the patient has it.

The problem with this is that when other doctors assess the same patient, they only agree with the first doctor 60% of the time. Even more unsettling is the fact that when the same doctor went back and assessed the same patient just hours later, that doctor agreed with their original assessment just 65% of the time.

Thankfully, we have built a machine learning model that now performs on par with ophthalmologists. Within a year, it will outperform the ophthalmologists, leaving them with more time to do what they do best -- treating the patients with diabetic retinopathy rather than spending countless hours inefficiently screening for it.

How Can You Use Artificial Intelligence to Grow Your Business?

Many people are aware that the tech giants of the world take advantage of the technologies mentioned above.

Google uses deep learning to make self-driving cars more efficient at analyzing and reacting to what is going on around them, Netflix uses it to suggest shows, and Disney uses it to track moviegoers' facial reactions to see if they are bored or excited and to predict how future scenes will make them feel.

But how can you use these technologies to transform your business?

At Sisense, we use machine learning to create features that give you better access to your data, such as anomaly detection and chatBots. What if, instead of asking your Amazon Echo for the weather, you could ask it for yesterday's gross sales figure and receive an answer immediately. Or how about mid-conversation with your team on Slack, asking the Sisense chatBot to "summarize sales by region for the fourth quarter of 2016.” Instantly, the chatBot will present you with a written response and a graph to elaborate on the data provided.

Artificial intelligence is no longer too expensive for the everyday business owner. As AI gets cheaper and cheaper, its applications continue to grow.

Today, companies use artificial intelligence to become more efficient by giving their employees valuable information in real time so they can make smarter, faster decisions with greater confidence. This enables the company to test products, strategies, and hunches more efficiently than ever before.

I look forward to posting more about big data, artificial intelligence, and the technologies of the future. As always, drop a note if you have comments, questions, or topics that you would like to learn more about.