Constrained Markov Sequence Generation

This book deals with the problem of generating sequences in the style of a given corpus, in text or music. In particular, the authors study Markov models, which have long been used for capturing basic statistical information about how elements of a given alphabet are put together in representative sequences, formulating them as constraint satisfaction problems (CSPs), a formulation which opens the door to many extensions of basic Markov models through the modularity of constraint satisfaction.

The book will be valuable to practitioners, researchers, and graduate students engaged with algorithmic composition and constraint satisfaction.

• Algorithmic composition
• Random walk (RW)
• Markov chains
• Markov constraints
• Chord generation
• Melody generation
• Virtuoso melodies
• Content generation
• Allen relations
• Max order
• Plagiarism
• Plagiarism
• Belief propagation
• Palindromes
• N-grams
• Interactive composition
• Constraint satisfaction problem (CSP)
• Elementary Markov constraint (EMC)
• Variable-order Markov model (VMM)
• Domain consistency (DC)

• Laboratoire d'Informatique de Paris (LIP 6), Université Pierre et Marie Curie (UPMC), Paris, France

  François Pachet

• Laboratoire d'Informatique de Paris (LIP 6), Université Pierre et Marie Curie (UPMC), Paris, Ireland

  Alexandre Papadopoulos

• Sony Computer Science Laboratory (CSL), Paris, France

  Pierre Roy

Policies and ethics