Problem-solving and design thinking in Machine Learning and Artificial Intelligence
Design thinking is a channel for problem-solving. As humans, we apply design thinking in our everyday lives to solve problems we encounter. For example, humans have been able to find cures for epidemic outbreaks that have occurred over the years through thinking and scientific methods. This year 2020, we see the world experiencing the COVID-19 outbreak, and we all hope that researchers succeed in finding a cure soon.
In Artificial intelligence (AI), Machine Learning (ML) algorithms are applied to solve problems by the computer. ML engineers train the algorithms to understand problems using what computers can process: data. For ML engineers to create useful products that solve problems and make life easier, they have to apply design thinking to the process of development.
ML is an AI technique where the computer creates computational models through deductions from data analysis and model training to solve human-related problems. ML is machine-centred. The image below gives us a picture of the steps involved in ML and how it can be connected to design thinking.
Design thinking can be seen as the human-centred approach to problem-solving. Both design thinking and computational thinking (ML) for problem-solving can be outlined in five steps. I’ll explain them and tell you how they impact each other.
Empathize and Analyze
In design thinking, for you to come up with products that suit your users’ needs perfectly, you have to understand your users, try to see things from their point of view and understand what matters to them. This will help you analyze and compartmentalize their needs into parts that can be used as data by computers for decision-making.
Define and Synthesize
‘Defining’ in design thinking means stating what the problem is. It can be done by asking the 5 ‘WHYs’ which help you come to a conclusion of what the root cause of the problem is.
Synthesizing involves taking different parts of the analyzed data and forming a whole problem that is understandable to the machine.
While getting to know the stages involved in design thinking and ML for problem-solving let me remind you that they don’t always run concurrently like this and the processes mustn’t always follow a linear path.
Ideate
Brainstorming and creative thinking are what this stage entails. Come up with wild ideas and different creative solutions to the problem that has been defined. Come up with as many solutions as you can, there can hardly be too many ideas.
When you have been able to think up solutions that are feasible, you need to think of features for these solutions and the product requirement info.
Prototype and Tuning
Create a simple draft of the model to show how it works to solve the problem. A prototype should be fast and inexpensive to make and should be able to exhibit the working principle behind the solution.
The model can then be tuned and tweaked as the project progresses and necessary changes can be made to obtain optimal functioning.
Test and Validate
Your prototype has to be put to test to ascertain if it is a feasible solution to the problem posed at the beginning. After the test stage, you can always go back to other stages to tweak your solution until you get what your user truly needs.
Design thinking is an important part of the thinking process that we apply to solve the problems that our users encounter. One must be careful so as to not create products that do not meet the user's needs or do not solve any real problem.
