Rasa Algorithm Whiteboard - Understanding Word Embeddings 3: GloVe

In this video we're going to explore GloVe embeddings while also understanding why word embeddings are genuinely hard to understand. We'll train our own variant of these embeddings too and you can find the code for this on github as well as colab. This material can be found here;
github.com/Ras...

[@r5bc]

I saw the master class play list and i saw all the videos in this this one too. And i want to thank you for the amazing work, the extremely big value you provide to us and the light you shed on the topic . I can't wait to see the next video. Please keep up the good work 👍👍👍👍

[@duongphanai7094]

this video saved my life! Good job! Keep it up!

[@firstnamelastname3106]

bruh, you somehow managed to explain it better than the dude from stenford, thanks !

[@WahranRai]

10:01 I think it is - log(Cij) instead of + log(Cij) (similar to square root error)

[@faangsde]

One of the best channels/videos regarding NLP. You strike the balance between video length and in-depth explanation of every topic. Also, the video is easy to follow and instructive.

[@kajumilylys2617]

I would like to know why do the weights change every time we retrain the embedding layer!?

[@SreeramAjay]

at 9:56 , should that be (minus) -log(cij) ?

[@masoudparpanchi505]

🤔🤔

[@a7d2e5aff]

Merci!

[@abdulrahmanmohamed2824]

Great, explnation! but what about if there another one like this for BERT?

[@RasaHQ]

A few videos later in the series are about transformers/attention which are the core components of BERT.

[@amirhosseinramazani757]

I really enjoyed it! Thanks!

[@kasraamanat5453]

that was so good, thank u❤

[@PieroSavastano]

Beautiful!

[@anupammitra]

Would like to thank Rasa for amazing work on explaining the concepts. Truly exceptional

[@AttiDavidson]

Thank you very much! Very good explanation.

[@kajumilylys2617]

Thank you for this series it's been super helpful ❤️

[@masoudparpanchi505]

well explained

[@anupammitra]

What are the publicly available word embedding available for use apart from GloVe

[@RasaHQ]

Hi Anupam! There are a lot of different word embeddings available. Some of the most common ones are Glove, Fasttext and word2vec. Contextual embeddings are also increasingly popular, including Elmo and BERT. (All of these are for English; embeddings are also available in other languages.)

[@anupammitra]

@@RasaHQ Thank you for your response