Re:Why do tree-based models still outperform deep learning o...
by Scene_Cast2 on Tuesday March 05, 2024 @09:43AM
(#39604062)
The team behind Yggdrasil tree library at Google was doing some interesting research into tree differentiability (and thus unlocking SGD & end-to-end learning for hybrid architectures).
[ Parent
| on HN ]
Re:Why do tree-based models still outperform deep learning o...
by mistrial9 on Tuesday March 05, 2024 @10:16AM
(#39604437)
No doubt that the research is good, but I find that name offensive, honestly
Re:Why do tree-based models still outperform deep learning o...
by tomrod on Tuesday March 05, 2024 @12:37PM
(#39606680)
I probably balance you out -- I love the name!
Re:Why do tree-based models still outperform deep learning o...
by queuebert on Tuesday March 05, 2024 @10:19AM
(#39604479)
Why?
Re:Why do tree-based models still outperform deep learning o...
by Rexxar on Wednesday March 06, 2024 @02:56AM
(#39613438)
Is it this library https://github.com/google/yggdrasil-decision-forests ?
Re:Why do tree-based models still outperform deep learning o...
by Scene_Cast2 on Wednesday March 06, 2024 @10:29PM
(#39624773)
Yep.
Re:Why do tree-based models still outperform deep learning o...
by Rexxar on Wednesday March 13, 2024 @05:49PM
(#39697986)
Thanks
Re:Why do tree-based models still outperform deep learning o...
by melondonkey on Wednesday March 06, 2024 @09:57PM
(#39624597)
This is interesting. Are BART models differentiable? I haven’t looked closely at them but I would have thought for posterior sampling they’d have to be. BART has been around for a while, too