Intro to Feature Engineering with TensorFlow - Machine Learning Recipes #9

Hey everyone! Here’s an intro to techniques you can use to represent your features - including Bucketing, Crossing, Hashing, and Embedding - and utilities TensorFlow provides to help. Also included is a walkthrough of using TensorFlow Estimators to classify structured data.

Links from the video:

Thanks, and have fun!

[gounna]

It is very short but every second matters. Seriously well designed well compressed contents. I appreciate Josh Gordon and Google!

[prof_shixo]

I think this should be transformed into a complete series in machine learning engineering. Please, try to develop such a curriculum as it would tremendously helpful not only for developers but also for researchers!

[MagicmathmandarinOrg]

Great job at explaining concepts in plain terms. John Gordon is a master at making hard concepts easy to even a nine year old child. Thank you!

[JoeWong]

Thanks for creating this series Josh Gordon you have increased my understanding of ML/AI greatly and I hope you keep this series going

[sudheerrao]

thanks Josh Gordon..learnt a lot from your videos...very much informative... hope you wil continue the series... waiting for next one

[ahmadallish]

the code link is leading to a 404 page

[techstuff]

Nicee I just watched all the videos until this episode.
And i must say it is the only ML course i can realy understans!!

[Jaggigee]

great job done by google, simplifying this by such great examples and awesome videos. Keep sending more please. Thank You!

[gyateen]

His plan is to complete the course by year 3017

[regivm]

It will be great if you could provide links to few reference materials for reading (or examples)

[yanghu]

Thank you. Looking forward to more on feature engineering.

[sifiso]

You are back, thanks for the tutorials

[shanker]

Thanks for such awesome videos :) The code link is not valid any more whats new link ?

[ekarl]

For the education example you say, "the best way to represent this, is just to use the raw value", but if there was a roughly linear relationship between increasing education category and your earning, could you not then re-code it as an ordinal numeric variable in order to save a few degrees of freedom / learn only a single parameter?

[swapnils]

Thanks for the valuable information, Please keep this videos going on.

[DavidAxelrodP]

Another great video! Is there an embedding feature in keras.perprocessing?

[ray]

What's difference between hashing and embedding

[ajaykumarbharaj]

Please upload more videos and include reinforcement learning

[FalahgsGate]

Great....thanks for these videos and good nformation

[sridhark]

ML made easy, Please have series related to internals of ML algorithms .