xyz100

9 ms·

1.
by xyz100 21d ago
Presumably there is distillation or similar being used to transfer from a larger model to a smaller one.
2.
by xyz100 28d ago
There are many jobs though, why this one? I don’t know, even with just an undergraduate degree I’m sure there are jobs that are spreadsheet manipulation. You could make the same case there?
3.
by xyz100 3mo ago
An adverbial clause at the front of a sentence should have a comma? Or is that not what it is?
4.
by xyz100 4mo ago
Do Apple Watches sufficiently track sleep apnea? Or is going in for a sleep study the only way to get a decent signal of if one is suffering from sleep apnea?
5.
by xyz100 4mo ago
It seems the case. I used to think weed was for enlightened individuals given how it was portrayed as a natural alternative to pharmaceuticals (e.g., for pain or psychological reasons). Many forms of media would discuss it positively. But i
6.
by xyz100 4mo ago
Alternatively, you win the highest prize in your field and the only path you see forward is solving the most impossible problem in the field, which of course you will fail to do.
7.
by xyz100 5mo ago
Since I am hitting the reply depth: You “solve” a dataset or task when you translate some model into actual real world problems by creating a model that actually “works” (not just high accuracy). What is otherwise the point of training the
8.
by xyz100 5mo ago
If you want to be literal with language, then do you ever really “solve” anything? Even tying your shoes is not solved. One day you may tie them better, but for practical purposes we can say it is solved. Likewise, you can spend 55 hours of
9.
by xyz100 5mo ago
MNIST (the number classification task) has been “solved” a billion times and it is hard to imagine any subsequent advances there as scores using a variety of methods have hit the saturation point of accuracy. Any further improvements are li
10.
by xyz100 6mo ago
You solve a dataset when you learn what there is to learn about the phenomenon of interest. The limit of such phenomenon is “cure all disease”, and clearly this is not solving that.
11.
by xyz100 6mo ago
What makes this dataset or problem worth solving compared to other health datasets? Would the results on this task be broadly useful to health?