NLP for Preserving the Quechua Language

on YouTube ↗31m

" summary

A project builds an NLP pipeline to support revitalizing Bolivian Quechua by ingesting four Bolivian PDFs into a vector store and running a vectorized RAG (retrieval-augmented generation) pipeline. The system is deployed using n8n, Google Gemini, and Streamlit, keeps a 10-conversation memory, does not fine-tune the model, and prioritizes Bolivian-dialect materials because official resources are scarce and often scanned; the demo works for some users but can be intermittently unavailable on free tiers.

AI-written summary · about

" chapters (7)

1. 0:00Quechua introduction
2. 6:48Bolivia challenges
3. 11:01Quechua bot with RAG N8N
4. 16:20Bot demo Streamlit
5. 19:59Low-code tools impact
6. 21:23Cultural preservation closing
7. 21:45Q&A

" related talks

• 11mNLP Spanish Challenges AlternativesIsabel Yepes · PyData Global 2020
• 23mMultilingual Natural Language ProcessingGajendra Deshpande · Conf42 Machine Learning 2021
• 40mThe indigenous languages technology project at NRC: An empowerment approach to languageRoland Kuhn · Big Data & AI Toronto 2020
• 30mBuilding Complete State of the art Natural Language ProcessingDavid Talby · Scale By the Bay 2021
• 37mBuilding Generative AI ApplicationsTuana Celik, Mia Chang, Dmitry Goryunov · AWS Summit Berlin 2023
• 41mAdding Generative AI to Real Time Streaming PipelinesTimothy J Spann · Build Stuff 2024
• 21mFrom 0 to a crowd sourced neural network in a jiffyJess Alfredsen, Sebastian Steenssøn · Data Science Conference 5.0
• 47mReal-Time AI with Open SourceTimothy J Spann · Big Data & RPA Conference Europe 2024