Studies in Computational Linguistics presents authoritative texts from an international team of leading computational linguists. The books range from the senior undergraduate textbook to the research level monograph and provide a showcase for a broad range of recent developments in the field. The series should be interesting reading for researchers and students alike involved at this interface of linguistics and computing.
Also available as eBook on:
Part I Analogy-based methods Skousen's analogical modelling algorithm: a comparison with lazy learning; Analogy, computation and linguistic theory; Constraints and preferences at work: an analogy-based approach to parsing grammatical relations. Part II Connectionist methods Towards a hybrid abstract generation system; Recurrent neural networks and natural language processing; Learning the semantics of aspect; Experiments in robust parsing with a guided propogation network; Learning semantic relationships and syntactic roles in a simple recurrent network. Part III Corpus-based methods A system for automating concordance line selection; A machine learning approach to anaphoric reference; Some methods for the extraction of bilingual terminology; Probabilistic part-of-speech tagging using decision trees; A new direction for sublanguage NLP; Efficient disambiguation by means of stochastic tree substitution grammars. Part IV Example-based machine translation Towards automatically aligning German compounds with English word groups; Corpus-based acquistion of transfer functions using psycholinguistic principles; A natural language translation neural network; Direct parse-tree translation in co-operation with the transfer method. Part V Statistical approaches Evaluating the information gain of probability based PP-disambiguation methods; Automatic error detection in part-of-speech tagging; A method of parsing English based on sentence form; A new approach to center tracking; Structuring raw discourse. Part VI Hybrid approaches Coarse-grained parallelism in natural language understanding: parsing as message passing; Parameterized message-passing for non-head-driven parsing; A stochastic government-binding parser; Evolutionary algorithms for dialogue optimization as an example of a hybrid nlp system; More for less: learning a wide covering grammar from a small training set. Part contents.