5h ago

Researcher Disputes Info Theory Limits on Synthetic Data Benefits

——0——
Original post
OPLuca Ambrogioni#1822Luca Ambrogioni@LUCAAMB

@andrewgwils I strongly disagree. It's a straightforward answer to why using synthetic data disagree boost a classified generally does not buy you much

10:05 PM · May 22, 2026 View on X
REPLYLuca Ambrogioni#1822Luca Ambrogioni@LUCAAMB

@andrewgwils I strongly disagree. For example, it gives a straightforward answer to why using synthetic data to boost a classified generally does not buy you much

Andrew Gordon WilsonAndrew Gordon Wilson@andrewgwils

Just like the no free lunch theorems, the data processing inequality (DPI) is irrelevant to the practice of machine learning, and that’s okay. It doesn’t mean we can’t develop info theory to be descriptive of practice.

11:16 PM · May 22, 2026 · 6.5K Views
5:12 AM · May 23, 2026 · 122 Views