Cognitive Procurement — A Glossary of Terms

Jordan Early
Cognitive Procurement
6 min readSep 25, 2016

Word association time…

What word comes to mind first when some says procurement? Costs savings? Business partner? Policeman perhaps?

I doubt very much whether ‘bot’, ‘machine learning’, ‘cognition’ or ‘artificial intelligence’ jump to mind. But the truth is the term procurement is inextricably linked with these terms and over the coming months and years they are only going to become more closely tied.

Do you understand these terms and what they mean for procurement? Do you, like many others, worry that robots will take procurement jobs? Should we build a wall?

If you are confused, don’t worry, you are not alone. If you have unanswered questions about AI or cognitive procurement read on. The following glossary will inform you and allay your concerns.

Cognitive procurement

It pays to start at the start, so let’s begin with cognitive procurement. Mark Perera, founder of Old St Labs, Procurement Leaders and a self confessed procurement tech-head defined the term earlier this year. Let’s take a look.

Put simply “Cognitive Procurement” is the intersection of procurement and cognitive computing.

Cognitive computing

“Cognitive computing is the simulation of human thought processes in a computerized model. Cognitive computing involves self-learning systems that use data mining, pattern recognition and natural language processing to mimic the way the human brain works.”

Procurement

“Procurement is the act of acquiring, buying goods, services or works from an external source, often via a tendering or bid process”

Thus;

Cognitive Procurement

Cognitive Procurement is the application of self-learning systems that use data mining, pattern recognition and natural language process to mimic the human brain to around the processes of acquiring, buying goods, services or works from an external source.

Artificial Intelligence.

Now that cognitive procurement is out of the way, I’d like to attempt a turn back to a more traditional alphabetical ordering for the remaining terms. To define artificial intelligence, let’s turn to Silicon Valley and its most reputed bastion of tertiary education and knowledge, Stanford University who define AI in the following way:

It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.

Another way to define the term is as follows.

Developing intelligent machines and software that are capable of perceiving the environment and take action when required and can learn from these actions.

Pretty broad right? I think the best way to think of AI is as a catch all or umbrella term for anything to do with computers acting more intelligently or more humanly. It’s the broad nature of the term that leads to its identity crisis. Everything from self driving cars to Spotify recommendation engines to Siri is classified as AI, as this space continues to grow, so too will this definition continue to broaden.

Coincidentally, I asked Siri if she was built using artificial intelligence. Her responses can be seen adjacent. As you can see, there is room for improvement.

Deep Learning and Machine Learning.

Up next and once again messing with my desire to create an alphabetically ordered cognitive procurement glossary is machine learning. The organised, systemic, OCD part of my brain would like to lead with deep learning however, deep learning is a subset of a broader concept, that of machine learning so we’ll need to start there. I’ve lifted the below from an earlier blog post I wrote called What Machine Learning Is and Isn’t.

Machine learning represents the next step in the evolution of the artificial intelligence movement. Moving beyond if/then scenario planning, machine learning focusses on producing algorithms that, through processing huge amounts of data, can learn (if a computer improve its performance of a task based on information it has gained from a past experience, we can say it has learned). This learning can then be applied to complete simple tasks.

It’s important to note however that the outputs of machine learning require a high level of human analysis and interaction before they are useful. Machine learning can indeed provide you with interesting data that corresponds to a recent drop in productivity at your biscuit factory, but you’ll need someone who understands the numbers and knows where to look to explain it all to you if you hope to get any value out this analysis.

Machine learning is not capable of making decisions or solving complex business problems on its own. It is entirely reliant on humans interpreting the results the algorithms return. Perhaps even more critical than the time spent analysing the data that gets spat out, is the time required to prepare the data that gets feed into the process. As we dealt with in It’s not the size of your data that matters, it’s how you use it : Garbage data in will equal garbage data out.

So what then is deep learning? Deeplearning.net defines the term on its website in the following way.

Deep Learning is a new area of Machine Learning research, which has been introduced with the objective of moving Machine Learning closer to one of its original goals: Artificial Intelligence.

Which I would suggest that to the procurement brain is rather unhelpful. I prefer this (slightly) more tangible description

A type of machine learning that uses a layered, brain-like network of “neurons”

My personal definition would go a little like this.

Machine learning is a singular process. A one off improvement. Deep learning continues to add layers on top of that initial process. If machine learning can make a single improvement, deep learning adds another improvement on top of that and then another and another based on the outcomes of the first improvement. The simplest example of this the much discussed Go (a board game) simulator produced by Google. The Google product, driven by deep learning, is now capable of beating the best Go human players. Not because it makes a series of unrelated clever moves, but because it can understand the impact it’s current move will have on its moves into the future. Machine learning act, deep learning can strategise. Furthermore, AlphaGo, the google simulator, can learn other tasks.

“The IBM chess computer Deep Blue, which famously beat grandmaster Garry Kasparov in 1997, was explicitly programmed to win at the game. But AlphaGo was not preprogrammed to play Go: rather, it learned using a general-purpose algorithm that allowed it to interpret the game’s patterns…” nature.com

Natural Language Processing.

This is where AI starts to get smart and starts to interact with us, the humans.

Natural language processing is a field of computer science involving interactions between computers and human languages.

This is a critical element of the AI picture as only by completely understanding human language can computers begin to interact with us understanding our requests and expectations.

Structured and Unstructured Learning.

It sounds a bit like primary school doesn’t it? Well its not. This one is all about the data.

Supervised learning is when you are utilising labeled or structured data to try and produce a singular defined output. You are trying to solve a specific problem. You are looking for the correct output from each input. Structured data is probably the data that most of us are familiar with, its the stuff your ERP system spits out.

Unsupervised learning, on the other hand, involves analysis of data that not labeled and is considered unstructured. Unstructured data could be video, audio, website content, forums, google searches, customer complaints or social media sentiment. Typically, it’s data that is external to the your company, it’s the big wide world. When dealing this with this sort of data the end product or output is (generally speaking) not known. Your machine learning processes will be looking for patterns in the data that can then potentially be analyzed and acted up.

Is that enough? Is the fog clearing? Are there more terms we should add to the glossary? Let us know and we’ll update this as we go!

Liked what you read. Share it, people will think you’re intelligent and well read.

--

--

Jordan Early
Cognitive Procurement

Aussie in San Diego. Writing on procurement innovation and remote working.