Machine Learning – Building vs. Applying
John Hennen | 10/21/2020
It is probable that through your IT and business journey that you have encountered the term “Machine Learning”. It is also likely that you have a good grasp, even if only intuitive, on what it means and what it can do. One thing I have seen across the landscape, however, is a gap in understanding that there is a fundamental difference between BUILDING and APPLYING machine learning algorithms. Let us dig in!
What is the official definition of machine learning? It is the process of using mathematical models of data to help a computer learn without direct instruction. It is considered a subset of artificial intelligence. Machine learning uses algorithms to identify patterns within data, and those patterns are then used to create a data model that can make predictions. With increased data and experience, the results of machine learning are more accurate – much like how humans improve with more practice.
The below video talks about “Machine Teaching”. This is a technical approach, to be sure – but you can also consider it a mindset. We need to consider machine learning to be much more than just feeding data into a black-box and expecting the correct output. We have to evolve the models, based on business needs, to produce the output we desire.
As Cassie Kozyrkov of Google put it (paraphrasing): building the algorithms for a machine learning model is like building a microwave from scratch, whereas applying machine learning to solve business problems is like using the microwave to warm food as part of preparing a fancy dinner. This drives you toward the question – do you need to understand the nuts and bolts about how a microwave works to use it to help cook a fancy dinner? Or can you simply assume it will do what you expect and move on?
It would be fair to say you do not need to be able to build a microwave from scratch to cook dinner with it. However, you DO need to be aware of how to use it properly, what the expected “inputs” and “outputs” should be, and how to validate that it is operating as expected. The microwave will deliver what you ask for, not necessarily what you want. And this brings us back to machine learning, and how Tuatara can help.
Tuatara approaches machine learning from both angles – applying and building. We start by working with you to understand your business problems, what you are trying to accomplish, and what data is available for analysis. Our approach meets customers where they are, identifies top priorities that will drive revenue and cost savings, sets the desired end state, then creates an iterative plan to achieve that end state over time. That plan will likely include building new data models or leveraging existing ones but will have the key input of a better understanding of what that data model is intended to show and explain, and how to best pull out the insights. Much like the microwave, if you do not understand how to tell a data model what to do, it may deliver less value than expected – or even incorrect information! We are here to ensure that does not happen.
Is this an opportunity within your company? Book a 30-minute meeting with our team to discuss how we can help your organization.
Director of Business Technology
John Hennen is the Director of Business Technology of Tuatara Consulting. He is responsible for the development and execution of Tuatara’s technology vision. He brings his deep customer-oriented business perspective to the table combined with the technical expertise required to guide his customers to their desired outcomes.