Learning Outcomes:
         ------------------
         1. The student understands the basics of the finite transducers and
         probabilistic weighted automata, and transition systems
         2. The student understands the crucial HMM algorithms
         3. The student can explain the relevance of machine learning and
         transition
         systems to parsing.

         Learning activities:
         --------------------
         1. Online: reading text book chapters
         2. Online: exercises related to the theory chapters
         3. Online: simple programming exercises (manipulation of HMM)
         4. Classroom where students make questions
         5. Seminar-type presentations

         Assessment:
         -----------
         1. 70% of exercises
         2. activity in the lessons (peer assessment)
         3. presentation

         Syllabus
         --------
         1. Fundamental structures
         relations, monoids, words, semirings, matrices, boolean circuits
         [programming exercise, e.g. matrix product in Python]
         [presentation on the history of automata and neural networks]

         2. Unweighted automata
         Kleene theorem, equivalence classes, nondeterminism
         [programming exercise]
         [presentation on engineering of automata]

         3. HMM
         forward algorithm, Viterbi algorithm
         [programming exercise]
         [presentation on POS tagging approaches]
         [presentation on codes and automata]

         4. HMM
         forward-backward algorithm
         [programming exercise]
         [presentation on learning of automata]

         5. Probabilistic automata
         weighted automata, transformations, HMM vs. PFA
         [presentation on applications of PFAs/HMMs]

         6. Rational transductions
         properties, equivalence properties, expressions
         [presentation on transducer algorithms]

         7. Context dependencies
         MSO, context conditions, maximum entropy models
         [presentation on transition systems in parsing]

         8. Further presentations
         [presentation on natural language complexity and the depth hypothesis]
         [presentation on registered automata or pushdown automata]

         Books (selected introductory parts):
         -----------------------------------
         - M. Lawson: Finite Automata.
         - J. Sakarovitch: Elements of Automata Theory.
         - K. Beesley and L. Karttunen: Finite-State Morphology
         - D. Jurafsky & J. Martin: Speech and Language Processing
         - R. Roche and Y. Schabes: Finite-State Natural Language Processing.
         - presentations rely on research papers