Increasing Variance Is Good in Deep Learning Networks

What these disruptive strategies do is they increase the variation in you're processing, often by introducing a lot of noise. What's good in variance? By opening up the variation, i pick up on bigger patterns that aren't changing,. That are real patterns in the world. This is what goes on in deep learning networks, right? In your attempts to get a grip on the world, there's good in right? There's good in Variance and bad in variants.