Information extraction 2400-ZEWW974

The course will sequentially cover the following topics:

1. Introduction to text data processing:
tokenisation,
stemming,
lemmatisation,
stopwords,
n-grams,
TF, IDF, and TF-IDF metrics.

2. Rule-based and heuristic methods:
pattern-based rules: pattern matching, regular expressions, Finite State Machines,
dictionary-based rules: exact dictionary matching, token-based matching, Aho-Corasick algorithm, Levenshtein distance,
syntax-based rules: dependency parsing, constituency parsing, context-free grammar, part-of-speech tagging.

3. Corpus-based methods:
TF-IDF,
RAKE,
topic models: LSA, NMF, PLSA, LDA, BERTopic.

4. Graph-based methods:
PageRank,
TextRank,
LexRank,
HITS.

5. Machine learning-based methods:
Hidden Markov Models,
Conditional Random Fields.

6. Neural network-based methods:
introduction to deep learning,
word and sentence embeddings, large language models,
recurrent neural networks,
LSTM, BiLSTM,
convolutional neural networks,
graph neural networks.

7. Transformer-based methods:
BERT,
GPT,
SpanBERT, LUKE, LLamaIndex, LangChain.

Szacunkowy nakład pracy studenta:
Typ aktywności K (kontaktowe) S (samodzielne)
wykład (zajęcia) 0 0
ćwiczenia (zajęcia) 30 30
egzamin 0 0
konsultacje 5 0
przygotowanie do ćwiczeń 0 10
przygotowanie do wykładów 0 0
przygotowanie do kolokwium 0 0
przygotowanie do egzaminu 0 0
… 0 0
Razem 35 40 = 75