Multilingual Natural Language Processing

on YouTube ↗23m

" summary

Shows a concrete pipeline that extracts text from images, audio and web pages using pytesseract / PIL, SpeechRecognition and Beautiful Soup, detects language, uses googletrans (unofficial, 15k char limit) to translate everything into one source language, runs spaCy/stanza-based tokenization/lemmatization/summary heuristics, then re-translates to a target language. Demonstrates a working code walkthrough for multi-source summary generation but notes missing features (NER for Hindi, uneven language support in polyglot/inltk) requiring hard-coded fixes.

AI-written summary · about

" chapters (10)

1. 0:00Multilingual NLP with Python
2. 3:27Challenges in multilingual NLP
3. 5:56Processing multilingual sources
4. 8:29Translation and speech libraries
5. 11:15Extracting text from audio/images
6. 13:58Web scraping with Beautiful Soup
7. 15:50Stanza for multilingual NLP
8. 19:02Indic NLP Toolkit
9. 21:32Polyglot for global support
10. 22:35Summary and future

" related talks

• 01h 39mNatural Language Processing (NLP) in Python - From Zero to HeroKeith Galli · PyCon US 2020
• 43mBreaking the language barrier: how do we quickly add multilanguage support in our AI application?Jaya Susan Mathew · NDC London 2021
• 48mIntroduction to Natural Language Processing with PythonBarbara Fusinska · NDC Minnesota 2018
• 45mTranscend Language Barriers with Custom Machine Learning APIs: um...Eléa Petton · Devoxx France 2024
• 17mNatural Language Processing in AutomationSweta Patra · DevConf.CZ 2023
• 36mCrash Course in Machine Learning for Natural Language ProcessingAllen O'neil · C# Corner Annual Conference 2022 - Hyderabad
• 01h 51mNatural Language Processing in PythonAlice Zhao · PyOhio 2018
• 44mTeaching Computers ABCs: A Quick Intro to Natural Language ProcessingLory Nunez · PyCon APAC 2018