The sense of accomplishment can boost your morale in multiple folds. Challenge yourself on anything small (or) big. If you succeed once, that will set the tone for your success.
Is there ways to regularize a model when trained with multi loss function? Model tends to be biased towards one loss function.
Use of significance factor doesn't really work
It's so hard to control the mind ! It gets into trouble all the times knowingly. If there is a program to erase the troubling thoughts , more humans will be at peace !
99% of my advice from this tech talk still holds.
But in 2020, it makes more sense to publish your articles in a specialized publication / academic journal or on your own blog.
Both @TryGhost and @Hashnode DevBlogs are good tools for this.
You don't get better at writing just by reading.
You don't get better at thinking just by listening.
You don't get better at creating just by consuming.
Learning requires your mind to be active.
For instance, where you read in a foreign language, you maximally enjoy the act of reading when you're at that stage when you frequently encounter new words and you have to intensely concentrate to make sense of what you're reading. Learning itself brings pleasure
The growth that comes from having to face novelty & difficulty doesn't just have concrete practical benefits, it's also intrinsically enjoyable -- perhaps more than anything else