Shared Context Probabilistic Transducers
نویسندگان
چکیده
Recently a model for supervised learning of probabilistic transduc ers represented by su x trees was introduced However this algo rithm tends to build very large trees requiring very large amounts of computer memory In this paper we propose a new more com pact transducer model in which one shares the parameters of distri butions associated to contexts yielding similar conditional output distributions We illustrate the advantages of the proposed algo rithm with comparative experiments on inducing a noun phrase recognizer
منابع مشابه
Parsing Morphologically Complex Words
We present a method for probabilistic parsing of German words. Our approach uses a morphological analyzer based on weighted finitestate transducers to segment words into lexical units and a probabilistic context free grammar trained on a manually created set of word trees for the parsing step.
متن کاملContext-dependent probabilistic hierarchical sublexical modelling using finite state transducers
This paper describes a unified architecture for integrating sub-lexical models with speech recognition, and a layered framework for context-dependent probabilistic hierarchical sublexical modelling. Previous work [1, 2, 3] has demonstrated the effectiveness of sub-lexical modelling using a core context-free grammar (CFG) augmented with context-dependent probabilistic models. Our major motivatio...
متن کاملStudying impressive parameters on the performance of Persian probabilistic context free grammar parser
In linguistics, a tree bank is a parsed text corpus that annotates syntactic or semantic sentence structure. The exploitation of tree bank data has been important ever since the first large-scale tree bank, The Penn Treebank, was published. However, although originating in computational linguistics, the value of tree bank is becoming more widely appreciated in linguistics research as a whole. F...
متن کاملAn Overview of Probabilistic Tree Transducers for Natural Language Processing
Probabilistic finite-state string transducers (FSTs) are extremely popular in natural language processing, due to powerful generic methods for applying, composing, and learning them. Unfortunately, FSTs are not a good fit for much of the current work on probabilistic modeling for machine translation, summarization, paraphrasing, and language modeling. These methods operate directly on trees, ra...
متن کاملAlgorithmic Information Theory and Computational Complexity
We present examples where theorems on complexity of computation are proved using methods in algorithmic information theory. The first example is a non-effective construction of a language for which the size of any deterministic finite automaton exceeds the size of a probabilistic finite automaton with a bounded error exponentially. The second example refers to frequency computation. Frequency c...
متن کامل