The Machine Understands Text — Learning and Search at Micro and Macro Scale
summary
Compares concrete NLP tools and results: tokenization with NLTK and spaCy, TF–IDF and pretrained word embeddings for vectorization, and cosine similarity for recommendations and semantic search. Demonstrates scikit-learn notebook examples (an ~85% classifier and good F1s), uses pretrained word vectors and Hugging Face transformers for masking and summarization, and deploys search using scikit-learn pipelines plus Elasticsearch v7 vector support, arguing search engines outperform relational databases for efficient text search.
AI-written summary · about
chapters (8)
- 0:00Wprowadzenie do NLP i ML
- 7:31Podstawy przetwarzania tekstu
- 13:48Przygotowanie tekstu do analizy
- 20:41Wektoryzacja i klasyfikacja tekstu
- 25:41Rekomendacja i analiza sentymentu
- 29:55Sieci neuronowe w NLP
- 35:18Wyszukiwanie tekstowe i semantyczne
- 39:15Podsumowanie i wnioski
related talks
31mTaking AI/ML to the next level: Text!
35mIndustrial Applications of Machine Learning in Search
30mKick-Start your Understanding of Machine Learning with Python
13mA Beginner’s Guide to Semantic Search
30mSemantic Search and Neural Information Retrieval
41mResearch in the Age of AI
22mThe Capabilities of Semantic Search
25mSemantic Search