DEV Community

Aniket Satbhai
Aniket Satbhai

Posted on

Why ML over Rule-Based Systems

1.2

So, you decided to make your car drive autonomously. You start thinking about the rules that would be you think would be needed.

At a glance, you require to turn the vehicle right, left or keep it straight. You also need to adjust it's speed.

When you go at granular level, you would know that the depending upon road condition, terrain, surronding vehicles, road signs, signals, etc., the vehicle needs to know how much degrees to turn, how much to accelerate. It also needs to follow traffic rules of the region you need it to drive, avoid any possible collision, update the rules depending upon changes in the surroundings. You need to make sure the rules cover all the possible circumstances.

At microscopic level, you need to make more rules. Eventually, managing the number of rules seems impossible task.

With Machine Learning you can avoid this step.

Oldest comments (0)