As human beings we’re always trying to do a better job communicating, so it’s little surprise that we want our machines to do the same. In many ways machines do a much better job communicating with each other than we humans do. Getting two machines to communicate is fairly easy, as demonstrated billions of times a day across the Internet. You may experience packet loss from time to time, but data typically transfers without a hitch. Human beings on the other hand are always struggling to reach a greater understanding. If you can get someone to totally understand ten percent of what you’re saying, you’re a master communicator. The main challenge is that we can't communicate with machines in the same way they communicate with one another. Thankfully we are not like Neo in the Matrix; we don't have a port in the back of our heads that connects us directly to the network. For machines and humans to communicate, the machine must do a better job of existing in our world.
To enable machines to converse with humans, AI programs use natural language processing (NLP). If you’ve met Siri, Alexa or Cortana or used speech-to-text on your smartphone, you’ve already experienced NLP. As you speak, the computer identifies the words and phrases and appears to understand what you said. Natural language processing in digital media strategy makes the interaction much more human. For example, if you need a waffle recipe, you might google “best waffle recipe,” but with NLP, you could say something like “I'm cooking breakfast. Can you give me a good recipe for those big fluffy waffles like the waffles they serve at the Breakfast Nook?" Creating an expert system that could anticipate this question, figure out what the person wanted and deliver one or more recipes for great Belgian waffles would be impossible. An artificial neural network, however, could handle the job. It might pick out a few key words, such as “breakfast,” “recipe,” and “waffles,” and give you a recipe for waffles. Or it might focus on “Breakfast Nook” and rattle off a list of local restaurants. Or it might key in on “recipe” and “big fluffy waffle” and figure out that you wanted a recipe for Belgian waffles. It might even ask for clarification and respond with a question like “Would you like me to search for a Belgian waffle recipe?” If the neural network delivered the wrong answer, the person could simply say, “No, I need a recipe for Belgian waffles,” and the neural network would learn that the next time the two of you had a similar conversation, it should retrieve a recipe for Belgian waffles, and not a list of local restaurants. Many organizations interested in artificial intelligence also offer free communication services, and they use machine language to analyze conversations. They aren’t interested so much in analyzing what you say, but how you say it. Google has access to anonymized versions of your email and voicemail. Apple offers iMessage. Microsoft has Skype. These services give their AI programs a treasure trove of different types of human communications. They can use machine learning to identify patterns and draw conclusions about how humans use their natural language. Natural language processing isn't just about understanding the words; it's also about understanding the context and meaning. A few years ago one of the top Google searches was, "What is love?" At the time, when you plugged that question into Google you would get a long list of results. Most of them were about pairing rituals and the importance of feeling connected. This was the kind of response you'd expect from a network that's just matching keywords across a database. Natural language processing gives machines the ability to better understand the larger world. Now if you ask Siri or Alexa “What is love?” it understands that you're probably more interested in romantic notions of love, so you’re likely to get a range of more poetic and philosophical answers. And as you converse about love with these artificial neural networks, they’ll begin to learn more about what you really want to know. The Internet of Things (IoT) The Internet of Things (IoT) refers to the large and growing collection of everyday objects that connect to the Internet and to one another. These devices include smart thermostats that learn your daily habits and adjust automatically to keep you comfortable, smart watches that can track your daily activity and let you know when you’re meeting your fitness goals and smart refrigerators that can order groceries when supplies are running low. Some devices can even monitor your health and let you know when you need to see your doctor. Certain devices can also communicate with one other. For example, your smart watch can tell your smart locks to unlock the doors when you approach your home or turn on music when you enter your living room. As you can imagine, these smart devices generate massive amounts of data, and this data can be very valuable. As a result, many IoT companies invest heavily in AI programs. With AI, organizations can analyze this data to identify patterns that enable the organizations to react quickly to problems or to take advantage of emerging opportunities. An IoT car coupled with AI could monitor the car’s operation and identify patterns that indicate when the car needs to be taken in for service. You can now purchase an electrocardiogram (EKG) sensor that is nearly as accurate as the ones in your doctor's office. These devices can check your heart's electrical activity. They’re inexpensive enough that many companies are embedding them into smartphone cases and watches. Using AI, these companies can identify patterns in the data using unsupervised machine learning on a neural network. The network can review the EKG data of thousands or even millions of different participants to find patterns that might accurately predict if someone has an impending health issue. With these sensors, AI may not yet be able to discover breakthrough preventive measures or cures for serious illnesses, but they can identify patterns that provide doctors with additional insights to help them come up with preventive measures and cures. These devices can also notify you when your data matches an alarming pattern, so you know to check in with your doctor before your health condition becomes more serious. Remember that neural networks operate in a black box. No one except maybe the network really knows how the machine identifies patterns. We can create intelligent machines, but we still don't completely understand how they “think.”
0 Comments
Leave a Reply. |