When I start writing text messages or emails these days I automatically receive prompts for the next words that I may want to use, or even phrases to complete my sentence, as I’m typing.
The word and phrase suggestions are ones that are consistent with the type of language I would use. And the name suggestions are unique to people I know and communicate with regularly.
My Google mini recognizes family voices instantaneously and responds not only to our questions but adds custom comments directed at each of us individually.
It’s clear to me that my computer and smart phone and Google mini are learning what language I typically use. With increasing accuracy they can predict what I might want to write next based on words I am typing. And the voice activated devices are increasingly interactive.
How does this Machine Learning, or Deep Learning, actually work? And how will it shape Supply Chain now and in the future?
What is Deep Learning?
Conceptually I like the simple definition provided by mathworks.com:”Deep learning is a machine learning technique that teaches computers to do what comes naturally to humans: learn by example.”
That sounds very interesting but how is that possible? It turns out that with advances in computer algorithms, speed and capacity, and greater levels of data, the basic tools to facility this “learning” are there.
Wikipedia offers that “Deep learning is a class of machine learning algorithms that(pp199–200) uses multiple layers to progressively extract higher level features from the raw input. Most modern deep learning models are based on artificial neural networks. In deep learning, each level learns to transform its input data into a slightly more abstract and composite representation. The word “deep” in “deep learning refers to the number of layers through which the data is transformed.”
A distinction that is called out in the Wikipedia definition is that Deep Learning is different from Machine Learning.
Here we go to hackernoon.com for clarification.. They characterize the distinction between Machine Learning and Deep Learning as follows:
Machine Learning for dummies:
A subset of artificial intelligence involved with the creation of algorithms which can modify itself without human intervention to produce desired output- by feeding itself through structured data.
If you are interested in learning more, here is a good collection of machine learning courses to start with.
Deep Learning for dummies:
A subset of machine learning where algorithms are created and function similar to those in machine learning, but there are numerous layers of these algorithms- each providing a different interpretation to the data it feeds on. Such a network of algorithms are called artificial neural networks.
From this I surmise that some of the applications that we deal with every day could be derived from the deployment of either Machine Learning or Deep Learning technologies.
What are some examples where Deep Learning and Machine Learning technologies are being applied?
- Autonomous vehicles
- Fleet management
- Object identification
- Visioning systems
- Text/sentence creation
- Predictive analytics
- Voice recognition and response systems
- Predictive maintenance
- Warehouse/Distribution Centre management
- Fraud detection and prevention
- Cancer diagnosis
- Safety management
- Workflow management
Can Deep Learning, or Machine Learning, be Applied in Supply Chain?
A quick perusal of the list of some of the applications of Deep Learning and Machine Learning should immediately rings bells for Supply Chain professionals.
The Digital Supply Chain of the future will have Deep Learning, or Machine Learning, at the core of its strategic platform. Fuelled by the provision of Big Data and the end to end electronic connectivity of global Supply Chains Machine Learning and Deep Learning will be enabled for any and every desired Supply Chain application.
Anyone in Procurement, Logistics, Supply Chain, Finance, Operations and Planning knows that they are dealing with literally thousands, if not millions, of pieces of data every day. It is most often impossible to attempt to manually manage this data. And even when it is presented in the form of information it can still be daunting and overwhelming.
On top of that today’s Supply Chain is increasingly complex. Suppliers at all levels can be located anywhere in the world with multiple modes of transportation engaged at any time, myriad channels, and multiple processing points, and the proliferation of direct to customer deliveries and interaction.
As such having Machine Learning, or Deep Learning, capabilities embedded in the business processes that govern any operation within a company will be a huge strategic step forward.
The very idea of Artificial Intelligence may be intimidating for some. But the ability of Deep Learning and Machine learning technologies too dramatically improve the efficiency of any Supply Chain, or other Business function, is very real.
As opposed to being afraid of this technology it is important for Supply Chain and Business professionals to understand these technologies. With this understanding you are then in a position to define, shape and direct the strategic deployment of those technologies.
Further these enhanced capabilities will equip people with the future skills needed to lead and holistically run global Supply Chains better than ever before.
One thought on “What is Deep Learning? And What Does it Mean for Supply Chain?”
thanks for sharing such a information along with videos on Deep Learning that Mean for Supply Chain. Great post.
Comments are closed.