Deep Learning with applications in NLP

Deep Learning with Applications in NLP

The course takes a tour of the main topics in Natural Language Processing (NLP), presenting both traditional and deep learning algorithmic approaches for text processing and analysis.

The grade is obtained based on three components:

lab assignments that have to be presented during each lab in the first part of the semester (30%)
a project developed in the second part of the semester (30%)
a written test in week 15 of the semester (40%)

To graduate, each of the components above must be accomplished in proportion larger than 50%.

	Subject	Course	Lab	Readings/Useful links
w1	Intro in NLP, basic text processing (tokens, lemmas, stemming, edit distance, POS tagging) lexical databases, corpuses	slides	Practice with WordNet and NLTK
w2	Syntactic structure and dependency parsing	slides	Get familiar with syntactic and dependency parsing
w3	Language modeling: statistical approaches	slides	Implement an N-gram language model for the Romanian language
w4	Machine translation	slides	Machine translation - compute BLEU
w5 online	Topic Modeling	slides	Practice with LSA (SVD,NMF) and LDA for topic modeling	An introduction to Latent Semantic Analysis LSA web site David Blei's web page on topic modeling
w6	Text vectorization	slides	Visualizing word embeddings	Multilingual universal encoder
w7	Recurrent Neural Networks	slides
w9	RNN variants	slides	Project planning Project list	An Empirical Exploration of Recurrent Network Architectures
w10	Text classification	slides	Practice with (Multinomial) NB and (Multinomial) LR
w11	Transformers	slides	Project developement
w12	Transformers-part2	slides	Project development
w13	Explainability in NLP	slides	Project development
w14	Project presentations