GTÜElectronics Notes

Fundamentals of Machine Learning · Digital lecture note

Lecture 10 – NLP: Tokenization, TF-IDF, Word2Vec

Makine Öğrenmesinin Temelleri – NLP: Tokenization, TF-IDF, Word2Vec

Course
Fundamentals of Machine Learning (ELM472)
Semester
Semester 7 · Fall 2024–25
Format
PDF · 6 pages · 0.3 MB · handwriting-style LaTeX
Language
Turkish
Original note
Ders-10.pdf

Open the PDF (0.3 MB)LaTeX source (.tex)Original handwritten note

Lecture 10 – NLP: Tokenization, TF-IDF, Word2Vec – first page of the digital lecture note
Notice: The digital notes were created from the handwritten notes with the help of Claude (Anthropic); they are meant as a supplementary resource and may contain errors. Always rely on the original notes and the official course materials.

What is in this note?

Section headings of the note (in Turkish):

Text of the note

Show as plain text (Turkish; formulas and drawings are not included)

NLP (Natural language processing)

My token name is Mehmed. ?

Morfolojik gitmek istersek (şekle göre)

Benim adım Mehmet.

kelimelere bölerek word piece

→ Morfolojik gramere göre

Elimizde dokümanlar var

Tüm dokümanlarda iki kelime aynı frekansla geçmemeli

TF-IDF → baskınlığı azaltmak için

Continuous Bag of Words

Skip gram / n-gram

2 gram iki komşuluk

tersten de yapabilirim

Global vectors for word Representation (Glove)

Covid geldi yeni kelime

Benim adım Mehmet / Seq 2Seq

RNN → exploding vanishing unutma fazla

sonraki güncellemede LSTM

Bu metin kısa olmadıktan sonra context'i yakalayamıyor.

kelimeleri hidden layerle çarpıyoruz

hangi kelime ne kadar önemli

sonra dönmek yerine MLP koyuyoruz

2019BERT: Sadece Encoder , kelimeleri ayırabilirsin

BERT NLP'nin YOLO'su

BERT : Bidirectional Encoder Representations from Transformers

ELMO: Embeddings from language model

context mekanizması yok

← Lecture 9 – CNN, RNN/LSTM, encoding and TF-IDFLecture 11 – RNN, attention and reinforcement learning →

← All Fundamentals of Machine Learning digital notes