News for hackers. Stuff that matters.

A comment on Are deep neural networks dramatically overfitted? (2019). Parent comment

Re:Are deep neural networks dramatically overfitted? (2019) by jhgb on Monday April 05, 2021 @05:29AM (#26696940)
If I derive Kepler's laws from positions of planets, I have preserved elements of their motion for replay as well.
[ Parent | on HN ]
Re:Are deep neural networks dramatically overfitted? (2019) by Retric on Monday April 05, 2021 @08:38AM (#26698179)
The laws alone don’t have enough information to replay their motion without each planets location, a foci, the orbital period, and an associated timestamp or equivalent information.

So, include that data and it’s lossy compression, exclude it and it’s not.

[ Parent | on HN ]