AI, Machine Learning, and Deep Learning – a simple introduction for marketers

Every day the ways we connect with technology are more interactive, intelligent, and relevant. As technology becomes more cognitive and engaging, marketers need to be aware of the analytics behind this technology and how they can use it to better connect with their customers.

As consumers, all of us are engaging with technology in new ways. Sure, we have been able to talk to our phone, TV, or car for some time and give simple directions. But now these devices talk back and can engage in an iterative dialogue to perform more complex tasks, identify patterns in our behavior, and even make recommendations. And devices are taking in information from and making decisions based on their environment and performing autonomous activities like self-driving. 

This innovation is driving new capabilities like the robot delivery services in San Francisco, Redwood City, and later this year in London, Düsseldorf, and Bern. Robots delivering daily essentials and treats to your front door is one use case that demonstrates why Forrester predicts investment in AI and Machine Learning will triple this year.

But what is Machine learning, why now, and how will it improve marketing and how brands can connect to their customers?

Simply put Machine Learning enables computers to learn autonomously without being explicitly programed.

The recent explosion AI and Machine learning is driven by a bunch of cool new hardware and software -- better sensors, better data structures (Hadoop with distributed computing) and better processors (GPU’s) running more complex analytics. One of the key breakthroughs in machine learning that greatly improves how computers process and learn is deep learning.  Deep learning in turn is machine learning using neural networks.

So, what are neural networks? Simply put neural networks are learning systems modeled on the human brain comprised of multiple “neurons”.  Each network contains neurons, often lots and lots of neurons connected in layers. These neurons, like those in the human brain, take information, process that information and then make a decision to fire (pass information to other neurons) or not fire. Information is passed through the network of neurons and along the way patters can emerge. The combination of multiple neurons, working together, eventually lead to a decision. And like the human brain neural networks also change the importance or weight given to an input or pattern to make a decision as the network learns.  

In this example network each circle represents a neuron and the yellow and red circles represent what is known as the input and output layers in the network. The layers in between are known as hidden layers and the recent break-troughs that have enabled deep learning networks comes from the design and optimization techniques in the hidden layers.

Example deep learning networks are Restricted Boltzmann machines (good for feature extraction) Deep Belief Networks (good for prediction), and Convulsion Networks (good for image and voice processing).

Machine learning and deep learning techniques are becoming more prevalent. Banks use deep learning for fraud detection, Netflix, Amazon, Google, Facebook, and Spotify use deep learning to better understand consumers interests and needs the optimize content and advertising.  

Expect new uses as consumer wearables, connected homes and cars become more prevalent – pending data collection and anonymization techniques to protects consumers.

And this has huge impacts on marketing and marketers. For example, will booking systems and application forms become a thing of the past as they are replaced with an intuitive conversation with a chatbot? Or what happens to marketing when your washing machine decides which powder to buy because it knows which one washes the whitest, causes the lowest amount of wear and tear and uses the least energy, immediately ordering the optimal product off Amazon?

So expect to hear more about machine learning and deep learning and how these techniques enable brands to use data and analytics to better understand their customers and provide relevant campaigns.