Tree probabilities

• Tutorial about context free grammar PDF

Expectation Maximization

• Short EM tutorial with a good visualization of the algorithm for a simple coin toss toy experiment PDF
• Slides explaining Expectation maximization with multinomials as an example PDF
• Paper about how the EM-algorithm connects to the variational inference PDF
• Explanation of using Naive Bayes model together with ML and EM PDF

N-grams

• Kneser-Ney: State of the art smoothing wikipedia
• Stupid Backoff: Much easier to compute than Kneser-Ney: pdf
• Jurafsky and Martin: about N-grams, has a good overview and a section about Kneser-Ney PDF
• Methods for smoothing Website
• Good-Turing Estimate PFD
• OpenFST a framework for working with N-gram models Website

Variational Inference

• Slides gining an introduction to variational inference with multinomial mixed model as example PDF
• Original PhD-thesis presenting the variational inference method PDF

Wordnet - similarity and information content

• Resnik similarity, a measure to determine semantic similarity between synsets by information content of least common subsumer. PDF

Misc

• Embeddings on UD parsed data PDF
• Trainomatic
  Uses BableNet to generate a training/evaluation set for word sense disambiguation
  Website PDF
• Context2vec
  A project which uses a LSTM network to learn a vector-context for words in a sentence. Probably can be modified to train on UD-tagged data? PDF Github Modification of LSTM to work on tree structures PDF

Unix tips and tricks

• sort and uniq to generate ngrams link