Deep Learning Explained
March 19, 2016. A historic day as Lee Sedol, the world’s greatest Go player, tackles one of the toughest games of his life. His opponent? Google DeepMind’s artificial intelligence program, Alpha Go. The place? The Four Seasons Hotel in Seoul’s Gwanghwamun district a luxurious locale for a momentous match. 60 million viewers from China are transfixed. The English-language livestream brings in 100,000 people, on the edges of their seats. Watching the game in adjacent rooms, meanwhile, are a few hundred representatives of the world’s press, joined by expert commentators.
Deep learning is catching fire inside and outside of the tech community. But given the complexity of the topic, it’s hard for the non-tech community to understand what’s really going on behind the buzzwords.
I’m writing this article in an attempt to bring some clarity to the issue.
Let’s start with some definitions (from Wikipedia):
Artificial intelligence (AI) is the intelligence exhibited by machines or software. It is also the name of the academic field of study which studies how to create computers and computer software that are capable of intelligent behavior.
Machine learning explores the study and construction of algorithms that can learn from and make predictions on data.
Neural networks In machine learning artificial neural networks (ANNs) are a family of models inspired by biological neural networks (the central nervous systems of animals, in particular the brain) and are used to estimate or approximate functions that can depend on a large number of inputs and are generally unknown.
Deep learning is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data by using multiple processing layers.
For the non-techies, it’s sufficient to understand this concept:
Deep learning is a branch of machine learning that is showing really promising results by employing neural networks with many layers and parameters.
How did Alpha Go beat the world’s premier Go player? The answer lies in ‘deep learning’, a branch of machine learning which uses neural networks based on those in the human brain. It enables programs to learn like humans do, just by observing and interacting with the world. Just like a human being, AIs using deep learning will start to notice patterns over time and alter their behaviour to fit with them, becoming more intelligent in the process.
Deep learning was briefly popular in the 1980s before being rejected as bunkum by the AI establishment. But now that computing power has caught up with the effectiveness of algorithms, the tables have turned. Deep learning is now seen as the next great thing in the technology world. We are already witnessing bidding wars over the companies that research how companies can use deep learning, with Google a major player in attempting to buy up these firms.
If you are thinking of using machine learning and deep learning to modernize your business, ask yourself a few things:
- What kind of data do you have or could you collect?
- What decisions can be made and/or automated based on this data?
- Is it worth the effort to develop software to do so?