Artificial Intelligence (AI) is one of the most hyped concepts of the last few years. From self-driving cars and automatic speech recognition to machines that can learn to play video games, robots are taking over tasks that were traditionally deemed as an exclusive competence of humans. Most of these recent breakthroughs have been driven by a particular type of AI: Machine Learning.
Machine Learning is defined as a type of AI that focuses on the development of algorithms that have the ability to learn from data, find hidden relations between different variables and use those relations to predict the future outcome of events. While this is something a human can do, the scale at which modern algorithms can learn is beyond anything we’ve ever seen before.
In a media environment that has become increasingly complex, it makes total sense that Ad Tech is applying AI techniques to the challenge of optimizing advertising performance.
How does Billy Mobile use Machine Learning?
Billy Mobile has developed a proprietary technology, Active Bx, which is powered by Machine Learning. It has been specifically engineered to upscale results with great speed by matching relevant advertisers with high-traffic publishers.
Here, Active Bx is trained in the same way a child learns to walk. The first few times they would lose balance and fall down (negative feedback). However, with every attempt they will position their feet with more accuracy,coordinating better their movements until they can take their first two or three steps (positive feedback). This is an incremental process composed of several small improvements, but the learning curve eventually leads to walking -and even winning the 100 meter dash! Reinforcement learning algorithms receive a reward when they take the best possible action, and nothing if they choose poorly, improving themselves by selectively retaining outputs that maximise the received reward over time.
The challenge here is to control what the algorithms strive for. In the case of Active Bx, the only purpose of its existence is to maximize the publisher’s revenue by choosing the right ad for the right visitor.
Aday Robaina, PhD.
Head of Data Science at Billy Mobile
Posted: February 14, 2017