amitness / blog-comments

Repo to host the comments with https://utteranc.es/

https://amitness.com/

0 stars 0 forks source link

2020/05/zero-shot-text-classification/ #8

Closed utterances-bot closed 7 months ago

utterances-bot commented 4 years ago

Zero-Shot Learning for Text Classification

A summary of “Train Once, Test Anywhere” paper for zero-shot text classification

https://amitness.com/2020/05/zero-shot-text-classification/

kranthik13 commented 4 years ago

Interesting paper indeed!

Jivnesh commented 4 years ago

Nice summary

zolekode commented 4 years ago

awesome summary

danr-ccg commented 4 years ago

Interesting approach. It's interesting that they think that this is the first zero-shot text classification paper. The first one is actually: Ming-Wei Chang and Lev Ratinov and Dan Roth and Vivek Srikumar, Importance of Semantic Representation: Dataless Classification, AAAI (2008) and it was followed up by several others that used the term "dataless" rather than "zero-shot". If you only care about a NN approach, check out: Wenpeng Yin and Jamaal Hay and Dan Roth, Benchmarking Zero-shot Text Classification: Datasets, Evaluation, and Entailment Approach, EMNLP (2019)

BinhMinhs10 commented 4 years ago

i want ask you what application suitable apply zero shot??

terry07 commented 3 years ago

Great presentation! Can you provide us a github repo for this experiment? Thanks

yananchen1989 commented 3 years ago

Good notes. Looking forward further blogs about ZSL.

This post https://joeddav.github.io/blog/2020/05/29/ZSL.html also discusses the ZSL using NLI paradigm.

marouaghaouat commented 2 years ago

Thank you