What Advertisers Really Need to Know About Deep Learning

It’s not the all-purpose tool it’s made out to be

Humans have a predilection for calling every new technology a revolution. While some innovations do change industries, the reality is that most don’t make much of an impact. The advertising industry is confronted with these so-called game-changers nearly every day, so it’s important to understand what’s really going to transform your business and what’s going to cause unnecessary headaches.

Deep learning is one such technology, and it’s intricate, technical nature makes it especially difficult to assess. So, what is deep learning and how does it fit into your digital advertising efforts?

Let’s start with the basics. Deep learning is the newest member of the AI family and it’s being called the solution to complex prediction, relevancy and banner blindness. When reading articles about it in the news, you can get the feeling it’s a turnkey, one-size-fits-all solution to every problem that the digital advertising industry is facing.

You might be wondering how you could have missed an entire development cycle during which this absolutely crucial advertising technology was created, but there are two misconceptions in that thought process.

First, this isn’t a new topic. Researchers have spent more than 20 years on the subject of deep learning and have made significant progress in many domains, including image recognition and sound processing. Second, while the technology isn’t new, we’re just starting to realize the potential of deep learning in advertising, so you haven’t missed anything.

What is deep learning?

To understand deep learning, you first need to understand machine learning. There are three things to know.

First is supervised machine learning. It starts with a human definition, which the algorithm learns to recognize and categorize. A well-known example is a spam filter, which spots predefined features, like “free drugs” or “you’re the winner of a brand-new car.”

Supervised machine learning’s slightly more complex sibling is unsupervised machine learning. This starts with uncategorized data which the algorithm breaks down into groups by the interpretable pattern clusters that it recognizes. From there, it’s up to a person to interpret what the data means.

Deep learning, in theory, joins the best of those worlds. Not only do you not need to define the patterns it’s looking for, but you don’t need to explain what those patterns mean either. Deep learning can recognize that a picture of a dog shows a dog, without a person feeding its features into the machine upfront (supervised) or analyzing groups of interpretable features like legs and tails to find a meaningful data set (unsupervised).

The all-purpose tool         

The way we talk about deep learning often reminds me of TV commercials for all-purpose tools. They promise to solve a wide range of problems confronting a do-it-yourself community. Take the Swiss Army Knife, the prime example of all-purpose tools. You can open a bottle, fix a bicycle chain, insert a missing screw and much more. An all-purpose tool might be the best tool you own, but it’s not the only tool you need. When the problem is more complex, like building an entire cabinet, you need more robust tools that are purpose-built for the task.

Deep learning might just be the all-purpose tool we’ve been waiting for, but for complex scenarios like digital advertising, it can’t handle the whole job. For that, you still need purpose-built tools.

Making deep learning a digital advertising tool

Implementing a deep learning architecture in digital advertising would mean processing a lot more data than in other applications, such as image recognition, and doing so in real-time. In the case of image recognition, the system requires terabytes of data just to be trained on recognizing an image of a dog. Our advertising exchanges, however, are driven by programmatic buying technologies that are under much stricter latency constraints than other use cases for deep learning—single-digit milliseconds at most.

Right now, the immense computing power needed to pull off a true deep learning architecture in digital advertising is not feasible. While a deep learning model might be able to reciprocate hand-crafted features, such as the time since a user’s last visit on a retailer’s website, it’s not the best use of processing resources. Instead, we see deep learning models being used to prepare in advance the information required for real-time decision-making.

But putting all that aside, is it any better than traditional models? Not necessarily.

Different, not better

It’s not that deep learning is better than traditional machine learning or vice-versa, but you need to consider your objective. Deep learning will absolutely affect advertising performance in the future, but only in the context of the whole machine learning spectrum.

The important thing is recognizing that every tool has its use. A screwdriver is not better than a hammer, because they’re tools built for different tasks. To understand how deep learning can help your business, start by following the scientific method and running experiments on your own data and KPIs. Measure which solution benefits you more, regardless of its internal implementation or the hype you read.


Romain Lerallut is VP and head of the Criteo AI Lab, in charge of developing the uses of AI in Criteo products. Before the launch of the lab in 2018, he was a director in the engineering department, responsible for the development of large-scale machine learning algorithms applied to problems such as product recommendation or banner graphic optimization. Prior to joining Criteo, he taught computers to read handwriting at A2iA.