Decision forests in TensorFlow | Session

preview_player
Показать описание
Neural networks are everywhere these days, but they're not the only type of model you should consider when you're getting started with machine learning. Tree-based models like decision trees and random forests are much simpler to train and easier to interpret. In this Session, we introduce TensorFlow Decision Forests, a new library to help you train and interpret tree-based models using familiar Keras API inside TensorFlow.

Speaker(s): Josh Gordon, Mathieu Guillame-bert

Watch more:

#GoogleIO #ML #AI

product: TensorFlow - General; event: Google I/O 2021; fullname: Josh Gordon, Mathieu Guillame-bert; re_ty: Premiere;
Рекомендации по теме
Комментарии
Автор

This is really cool! Random Forests are one of my favourite kinds of models

mrdbourke
Автор

This is one of the most succinct tensorflow videos i've found.

dr_flunks
Автор

Thank you guys for this videos. Never used decision tree before. I can see you can build powerful pipelines using decision forest/CNN

dwebdesignsoftwareagency
Автор

Is it possible to join tree with custom models?
For eg autoencoders + randomforrest

hrashikeshtiwari
Автор

What happens to sklearn now? 🤔
Sklearn is wonderful library, not sure how exhaustive is keras as of now. But I'm sure tough competition for sklearn.

satty
Автор

I think sklearn will and still will have its position in machine learning, just look at its beautiful documentation.

tubeathrun
Автор

how do I know which color belong to which species?

ate
Автор

didn’t know I wanted this one, but glad it’s here. Love you Google xx

flightrisk
Автор

thanks very straight.. to the point and helpful

matashree
Автор

Thanks guys, I think it's really useful

vasylcf
Автор

How can we do ranking tasks using this API?

ravitp
Автор

it doesn't work on jupyter right or I just don't know how to install it.

LongTail
Автор

📺💬 You can create random forest networks from Tensorflow easy example you had input as .csv and you need to explore their possibility or priority example word inputs or possible options for a category before you build networks or you can use tree models as decision tree networks.
📺💬 What is my expectation from the tree model I want to excite examples? 🥺💬 Are you talking about snooker algorithms ⁉
📺💬 It is faster when working with priority and orders.

Jirayu.Kaewprateep