Can ML replace all decision logic?
Traditionally, software developers used logical gates or decision trees to make choices in code. When the criteria are met, the system can make a decision without human input.
One of the uses for Machine Learning is to allow the software to make decisions without hard-coded logic or semantic-based algorithms. Supervised learning is accomplished by teaching the systems which choices to make depending on various parameters.
In an environment where quick or complex decisions must be made, ML has a clear advantage.
What if Machine Learning was used to decide when to turn on the light after hitting a switch? Is it possible for a computer to decide when an action should be taken on human input?
Let’s use a light switch as an example. If we want the light turned on, we will press the button and expect it to work. Unless, of course, it was done by mistake. Wouldn’t it be great if the computer decided that no action should be taken and did not turn the light on?
What about applying this logic to transactions, communications, education, and other areas where decisions are typically made by humans? A lot of errors can be eliminated as a result.
Is this the future of ML?