3 ms·
Also, a lot of empirical progress in ML didn't happen until people stopped worrying about explainability and theoretical guarantees. Remember when NNs were unfa
by bravura 4y ago
Also, a lot of empirical progress in ML didn't happen until people stopped worrying about explainability and theoretical guarantees. Remember when NNs were unfashionable?
- dredmorbius 4y agoSelf-imposed unnecessary limitations are a frequent inhibitor of technical progress. Though I've had my concerns for what the growth in solution-without-explanation (or understanding) that ML is generating.
- NateEag 4y agomy impression is that mostly NNs made progress when compute got really cheap.