Algorithms have inputs and outputs.
But unless we can define our desired output, there’s no point feeding any inputs into the system.
Without clear questions, we can’t build reliable algorithms to solve our problems.
Attributes of well-defined problems include:
1) Clarity. Think “reduce LDL cholesterol to 2mmol/L” rather than “improve heart health.”
2) Focus. Think “improve liver function” rather than “improve lung, heart, and liver function.”
This might seem limiting.
But depending on the downstream applications of the algorithm, it doesn’t have to be (e.g. machine learning algorithms).
Next up: how do these algorithms behave when we feed them lots of stuff at once?