oh, this made me realize that sample complexity reduction is fundamentally forward looking, as the lossy compression is a tacit set of predictions about what is worthwhile. i.e. guesses about sensitivity of your model. Interesting.
oh, this made me realize that sample complexity reduction is fundamentally forward looking, as the lossy compression is a tacit set of predictions about what is worthwhile. i.e. guesses about sensitivity of your model. Interesting.