The concept of Artificial Intelligence (AI) has been around since the 1950s, but the use of AI to improve warehouse and DC operations is still in its infancy. Nevertheless, AI is expected to have a growing role in warehouse management over the next five years. This article answers the question What is Machine Learning, and is intended to help operations, engineering and IT executives understand what they need to know about this emerging technology.
How Machine Learning Relates To AI
Artificial Intelligence is a broad topic that is receiving tremendous amounts of hype, leading to abuse and misuse in marketing. According to one study, forty percent of European AI startups don’t actually use AI.
Putting aside the differences between marketing and reality, an AI expert from Carnegie Mellon University notes that “AI is a moving target.” Fifty years ago, a chess-playing computer was cutting-edge AI. Not so today.
Most of today’s cutting-edge applications of AI use machine learning, a form of AI in which learning algorithms are applied to large sets of data to create predictive models relating to specific business outcomes. Popular machine learning techniques include deep learning and deep neural networks – when you hear those terms, think “machine learning.”
How Is ML Different?
The thing that makes machine learning so compelling is that the ML models are not developed or maintained by teams of engineers and ML systems do not require explicit programming.
An example of explicit programming is software code that follows specific steps to determine what move to make on a chess board, or how to calculate the time it should take to perform a given activity. ML systems apply algorithms to existing data to come up with their models and answers.
Machine learning is widely used for things like facial recognition, speech recognition and email spam detection. It is also used for tasks like video suggestions on Netflix, or shopping suggestions on Amazon. In the supply chain space, ML is being used in multi-site inventory planning by predicting consumer demand. And robotics systems use machine learning to master complex tasks like navigating a warehouse.
Machine Learning Needs The Right Data
Machine learning requires large amounts of fine-grained data and tremendous amounts of computing power to process the data. But it’s not just about having lots of data.
To be effective, ML requires the right data for the questions it is intended to answer. For DC applications, that data would not usually be found in enterprise software systems that capture general transaction data, such as an ERP or WMS.
Instead, ML relies on streams of fine-grained data that is often associated with IoT (the Internet of Things) devices. IoT refers to interconnected machines (conveyors or sorters with sensors, etc.) and mobile devices that collect and share massive amounts of real-time data. For example, mobile devices used in RF or voice picking applications often collect time-stamped data about every user interaction with the system in addition to other data. In the past, some of this data may have been used for short term purposes (debugging, training, etc.), but it was not usually collected or saved for other uses. The data had no value beyond those immediate needs. But machine learning changes that by automatically discerning patterns and finding meaningful information buried in the wealth of IoT data that many DCs already have.