An algorithm is a set of instructions. A cookie recipe, software, or a list of directions to get to the post office are all algorithms. Computers are really good at following algorithms and have surpassed humans at completing them, but the concept itself is about as old as civilization. Where Do We See It Showing Up in the Marketing Industry? Algorithms show up lots of places because the word is trendy, but it’s true that we use instructions for just about everything anyway. It’s accurate but maybe a little silly to call a drink recipe an algorithm. In the many times technology marketers use it, it’s not necessarily out of place unless it’s used as an equivalent for AI, machine learning, or deep learning. An algorithm can be used to, for example, “take behavioral data from tens of thousands of email interactions in order to determine the optimal time to send an email where people are most likely to open it and follow the link inside,” says Maurizio Pittau. What is AI (Artificial Intelligence)? Artificial intelligence is basically computers imitating how humans think. This includes computational systems designed for just one specific purpose as well as software that’s truly designed as a general imitation of human thinking. The former category is called “weak AI.” Whenever AI is developed to think through problems without help and look for novel solutions like humans do, that’s “strong AI.” In other words, strong AI robots would be the ones to overthrow us in 2120. Today we only have Type 1 and Type 2 AI, which are “reactive machines” or “limited memory” systems at best. The best AI we have right now are systems that use past experiences to shape future decisions--i.e., self-driving cars. Of course, in spite of pop culture references to AI replacing human workers or becoming our computational overlords, today’s AI isn’t there yet. It may never be. TechTarget notes that Michigan State University’s Arend Hintze’s four categories of AI suggest that the ability of computers to have feelings and demonstrate self-awareness are likely many years away. How Do We See It Showing Up in the Marketing Industry? AI is occasionally used as a synonym for technology. SaaS products often get “AI” tacked onto them even if the closest they get to artificial intelligence used as part of a computer application. AI also gets conflated with data, analytics, and data analytics. AI’s core competency is its ability to think and learn. If a consumer is a pet owner, for example, Amazon or another company is able, through AI, to learn about what types of pets you have, what kind of food they tend to eat, and how often you buy it. The basics. What is Machine Learning?Simply put, machine learning is a special type of algorithm that uses statistics to make a prediction about data and become better at doing this over time. To test an algorithm and prep it for use, data scientists provide a large number of examples for it to “look” at. An algorithm designed to find cat pictures, for instance, could be given several million pictures of cats and pictures of things that aren’t cats. The algorithm makes its choices and receives feedback on how well it did. This essentially trains the system to make predictions. As the machine learns, it gets better at its job. How is Machine Learning Different Than AI? It almost looks like AI, but since we’re getting picky and technical here, machine learning isn’t really intelligent. It’s still highly dependent on the intelligence of human creators. Machine learning, says Dhruv Shah, “is the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something in the world.” As MIT scientist Luis Perez-Breva told ZDNet’s Greg Nicols, machine learning can help an algorithm sort market data to make stock predictions, but you can’t ask the algorithm directly “How might I actually make money?” How Do We See It Showing Up in the Marketing Industry? Some marketers conflate AI and machine learning. Machine learning is also applied to any use of data, even for systems that don’t actually learn. Worse, machine learning as a term can also be thrown into content for SEO purposes. In the case of our pet owner above, machine learning is what may tell Alexa to remind you that it’s time to order cat food, or make some other suggestion based on what it has learned about you combined with its ability to predict what you’ll want to do next. What is Deep Learning? Deep learning, like machine learning, takes data and becomes more accurate at making decisions with it over time. Unlike machine learning, deep learning does this without the close supervision of a data scientist. No one’s providing all the rules, the machine has to create its own rules to develop predictions. Deep learning systems have to draw their own conclusions about the data they’re given, learning in a similar fashion to how human brains learn. How is Deep Learning Different Than the AI/ML/Algorithms? Deep learning is similar to machine learning and is a use of AI, but it’s “smarter” than basic machine learning and not all AI applications necessarily focus on deep learning. Deep learning uses and creates algorithms. Where Do We See it Showing Up in the Marketing Industry? For the most part, deep learning shows up whenever someone’s talking about data. It’s true that deep learning can be used to process data and make it useful, but not every system using data is using deep learning. For marketers, Michael Brenner says deep learning can help with “content generation, real-time bidding on ad networks, chatbots, speech recognition, and natural language processing.” Pet owners of the world may find themselves viewing programmatic ads with a picture of a chihuahua if that’s the breed they own, or a Siamese cat of they’re more of the feline persuasion. No generic stock art; deep learning knows you and how to appeal to your sense of individuality.For the most part, deep learning shows up whenever someone’s talking about data. It’s true that deep learning can be used to process data and make it useful, but not every system using data is using deep learning. For marketers, Michael Brenner says deep learning can help with “content generation, real-time bidding on ad networks, chatbots, speech recognition, and natural language processing.” Pet owners of the world may find themselves viewing programmatic ads with a picture of a chihuahua if that’s the breed they own, or a Siamese cat of they’re more of the feline persuasion. No generic stock art; deep learning knows you and how to appeal to your sense of individuality. Why All The Confusion? These words and the ideas they represent are steadily gaining traction in both our economy and in marketing itself. We all hear them all the time anymore, whether they’re used correctly or not. Marketers want to follow trends and earn buzz--these are legitimate uses as long as they don’t make technology terms meaningless.
In marketing, we can get ahead of ourselves in excitement about the next new thing. We’ll keep talking up the features and characteristics of cool new tech. Getting the words right, though, keeps it honest and comprehensible. Humans are reading this, after all. |
SIGN UP!Maureen JannI'm a veteran digital marketer whose career has grown up with the Internet. 
Archives
October 2019
Categories
All
|
RSS Feed
|
Super Deluxe Marketing
Marketing Strategy & Graphic Design Maureen Jann Proud members of the
|
