Data Augmentation in NLP - kzbin.info/aero/PLsAqq9lZFOtUg63g_95OuV-R2GhV1UiIZ Research paper walkthrough- kzbin.info/aero/PLsAqq9lZFOtWUz1WEoJ3GXw197LD7BxMc
@linminhtoo3 жыл бұрын
You clearly know a lot and your explanation is very helpful! A suggestion for improvement would be to sound more enthusiastic / less monotonic when you're explaining! So that it's easier for us, viewers, to follow also (otherwise we get sleepyyyyyyy)
@TechVizTheDataScienceGuy3 жыл бұрын
Thank you Lin Min Htoo for the appreciation and relevant feedback. I will work on that definitely. Cheers!
@singhRajshree3 жыл бұрын
what about code-mixed language?? i wanted to balance my dataset with text data augmentation... any idea?
@TechVizTheDataScienceGuy3 жыл бұрын
I guess, all the methods mentioned in the paper are mostly language agnostic but do require some language specific toolsets (tokenizer, etc). So, for codemixed data you can utilise tool sets for languages that are mixed and inject noise patterns for both.