Ksenia's Public Channel on #whatever

Pretty cool!

this paper's nuts. for sentence classification on out-of-domain datasets, all neural (Transformer or not) approaches lose to good old kNN on representations generated by.... gzip (c)

It's super elegant. If x1 and x2 are similar, then their joint compression (lzw/whatever) is more efficient. (c)

https://twitter.com/giffmana/status/1679399205570326528?s=46&t=tJA3XyJ6UXEb-9Si1PfeJQ

www.tg-me.com/jp/Ksenia& 39;s Public Channel on whatever/com.ksenia_and_ai/544

91 viewsedited  Jul 13, 2023 at 19:19

Pretty cool!

this paper's nuts. for sentence classification on out-of-domain datasets, all neural (Transformer or not) approaches lose to good old kNN on representations generated by.... gzip (c)

It's super elegant. If x1 and x2 are similar, then their joint compression (lzw/whatever) is more efficient. (c)

https://twitter.com/giffmana/status/1679399205570326528?s=46&t=tJA3XyJ6UXEb-9Si1PfeJQ

BY Ksenia's Public Channel on #whatever