N-gram Grammars – Meta-Guide.com

Notes:

An N-gram grammar is a type of probabilistic language model that is used in natural language processing (NLP) to predict the likelihood of a sequence of words or tokens. The term “N-gram” refers to a contiguous sequence of N items, where N is a positive integer. In the context of language modeling, N-grams are used to model sequences of words or tokens in a text, with common values of N being 2 (bigrams) or 3 (trigrams).

An N-gram grammar typically uses a set of N-grams as the basic building blocks to model a language, and estimates the probability of each word given the previous N-1 words in the sequence, in order to predict the likelihood of a sequence of words or tokens.

It is used in several tasks like machine translation, text summarization, speech recognition, and text-to-speech systems. In these tasks, N-grams are used to estimate the probability of a word given the previous n-1 words.

Also, N-grams can be used for text classification, for example, in the case of spam detection. In this case, the algorithm will examine the text and compare it with a pre-built database of n-grams commonly found in spam emails. If the email in question contains a high number of those n-grams, it is marked as spam.

The references below mention the use of an N-gram grammar as a technique for speech recognition in various systems. It is used as a likelihood model for sequences of n words to describe how words occur in sentences. It is used to generate grammars for natural language processing and automated call routing systems. It is also used for language modelling, and it is commonly used in conjunction with other techniques such as grammar-based language modelling and statistical language modelling. It is used to estimate speech grammars and can be represented using finite-state transducers. It is also used for phrase boundaries and accent determination and concept-to-speech. The text also states that it can robustly model a huge set of customer queries.

Kazakh Noun Phrase Extraction Based on N-gram and Rules G Altenbek… – Asian Language Processing (IALP), 2010 …, 2010 – ieeexplore.ieee.org … symbols, because it requires a very large database. Therefore, an approximation method can be used to solve this problem. We can use the “N-gram grammar model”. In formula (2), the conditional probability can be approximated … Related articles – All 2 versions

[PDF] Detecting Structural Irregularity in Electronic Dictionaries Using Language Modeling [PDF] from trojina.si P Rodrigues, D Zajic, D Doermann… – Proceedings of …, 2011 – trojina.si … without smoothing. Tree Unique Tokens 2-gram 3-gram 4-gram ENTRY 21 178 395 667 FORM 7 25 44 51 SENSE 22 183 384 628 Table 2: Unique Token and n-gram grammar counts at each tree level. This language model … Cited by 1 – View as HTML

Problems and Review of Statistical Parsing Language Model F Zhou, F Zhang… – Asian Language Processing (IALP), …, 2010 – ieeexplore.ieee.org … are also analyzed. A. Bigram Lexical Dependencies The bigram lexical dependencies language model proposed by John Collins combines the thoughts of dependency grammar and the statistical N-gram grammar. It adds the … Related articles – All 2 versions

Utterance Classification for Combination of Multiple Simple Dialog Systems SJ Hahm, A Ito, K Awano, M Ito… – Parallel and Distributed …, 2011 – ieeexplore.ieee.org … At first, the input utterance is recognized by speech recognizers using a large vocabulary N-gram grammar as well as finite state grammars (FSGs) automatically generated for frame-based dialog subsystems. Then the task classifier … Related articles – All 2 versions

[PDF] Probabilistic Analysis of Sindhi Word Prediction using N-Grams [PDF] from insipub.com JA Mahar… – Australian Journal of Basic and Applied …, 2011 – insipub.com … Table 2: The sample of Sindhi words bigrams. The experiments are performed on our developed Sindhi corpora, for testing we have randomly selected 15,000 sentences, 3000 sentence from each corpus type and then applied N-gram grammars. The backoff 1140 Page 5. Aust. … Related articles – View as HTML – All 2 versions

Speaker independent Urdu speech recognition using HMM J Ashraf, N Iqbal, NS Khattak… – Informatics and Systems ( …, 2010 – ieeexplore.ieee.org … 3) Language Model Sphinx-4 supports three types of language models. The Context-Free-Grammar (CGF) is usually used in speech applications based on command and control. The n-gram grammars are used for free speech form. … Cited by 2 – Related articles – All 5 versions

Development of a Toolkit for Spoken Dialog Systems with an Anthropomorphic Agent: Galatea [PDF] from hokudai.ac.jp K Katsurada, A Lee, T Kawahara… – … APSIPA ASC 2009 …, 2009 – eprints.lib.hokudai.ac.jp … connection constraints. In this version, an engine can use either an N-gram, grammar, or isolated word recognition using only a dictionary. Along with this modification, Julian, a grammar-based engine, was merged with Julius. In … Related articles – All 10 versions

A Combination of Statistical and Rule-Based Approach for Mongolian Lexical Analysis L Zhao, J Men, C Zhang, Q Liu… – Asian Language …, 2010 – ieeexplore.ieee.org … of a directed graph G defined: ( ) ( ) PG PT P t T PT t T = × × T present a candidate tree, there are three candidate trees in Figure 1. P(T) is a candidate tree’s entire generated probabilities which can be understand as traditional n-gram grammar expanded model: … Related articles – All 2 versions

Spoken Vehicle control RS Hernández, JF López, D Scarlatti… – EP Patent …, 2011 – freepatentsonline.com … There are several techniques for language modelling including grammar-based language modelling and statistical language modelling (N-gram). Grammar-based language modelling consists of defining all possible sentences that the system can recognise. … Related articles – Cached

[PDF] EVALITA 2009: Loquendo Spoken Dialog System [PDF] from psu.edu E Giraudo, P Baggia – Evaluation of NLP and Speech Tools for Italian …, 2009 – Citeseer … Pronunciation Lexicon Specification (PLS) Reusable Components Call Control XML (CCXML) Semantic Interpretation for Speech Recognition (SISR) N-gram Grammar ML Speech Recognition Grammar Spec. (SRGS) Natural Language Semantics ML VoiceXML 2.0 … Cited by 3 – Related articles – View as HTML – All 7 versions

Systems and methods for providing automated directory assistance using transcripts RM Schwartz, H Shu, J Makhoul… – US Patent 7,401,023, 2008 – Google Patents … The tran- scription module (310) obtains transcripts relating to direc- tory service requests. The speech grammar estimation mod- ule (330) creates an n-gram grammar for multiple telephone numbers from the transcripts. The … Cited by 1 – Related articles – All 2 versions

Study on question-answering system based on Meta search engine F Li, H Kang, Y Zhang… – Computer Application and …, 2010 – ieeexplore.ieee.org … These words not in the dictionary will be cut apart, so they will impact the precision of information retrieval and the precision of answer extraction. In this paper, new word recognition, which uses n-gram grammar model, is added to the information retrieval module. … Related articles

Adjacent reordering phrase-based translation models L Yujian, L Yanpeng – Natural Language Processing and …, 2010 – ieeexplore.ieee.org … Usually, Pr (t/ ), called the language model, IS processed by n-gram grammar, while Pr ( s; It/ ), the … tK =tl =tpt2, … ,t] is generated by a n-gram grammar of words in target language, namely, f Pr(t;K) == Pr(t{) == Pr(t])· IlPr(t; It;-]) ;2 n f = Pr(t])· ITPr(t; It;-]). …

Implementing a hybrid dialogue strategy for speech-enabled mobile commerce Y Fan – International Journal of Electronic Business, 2009 – Inderscience … The system uses N-Gram grammars to govern the recognition at the request segment of a dialogue, and employs an application-directed strategy at the clarification discourse segment. The paper also details generating a corpus … Cited by 1 – Related articles – All 7 versions

Learning long-distance phonotactics [PDF] from udel.edu J Heinz – Linguistic Inquiry, 2010 – MIT Press Page 1. Learning Long-Distance Phonotactics Jeffrey Heinz This article shows that specific properties of long-distance phonotactic patterns derived from consonantal harmony patterns (Hansson 2001, Rose and Walker 2004 … Cited by 20 – Related articles – All 10 versions

Natural language interface control system CC Konopka, LC Almstrand – US Patent 7,447,635, 2008 – Google Patents … HZ j (… I ( Feature Extraction N-gram Grammar Probabilistic Context Free Grammar Speech Decoder 1) ii Natural Language Interface Acoustic Models Device Interface Feedback Module Controlled V Device(s) Ti i I(tm)” 1[J|_a Device Abstraction(s) L System Processing Control … Cited by 9 – Related articles – All 2 versions

[BOOK] Inductive learning of phonotactic patterns [PDF] from ucla.edu JN Heinz – 2007 – books.google.com … 65 B-3.10 Languages and the Machines which Accept Them 65 B-3.11 Tail Canonical Acceptors 66 B-3.12 The Myhill-Nerode Theorem 67 B-3.13 Head Canonical Acceptors 68 3 Patterns over Contiguous Segments 76 1 Overview 76 2 N-gram Grammars and Languages 77 3 … Cited by 44 – Related articles – Library Search – All 8 versions

An FPGA implementation of speech recognition with weighted finite state transducers [PDF] from mirlab.org J Choi, K You… – Acoustics Speech and Signal …, 2010 – ieeexplore.ieee.org … Four hierarchical knowledge sources usually uti- lized in speech recognition can be represented by WFSTs: HMM structure H, context dependency C, pronunciation lexicon L, and n-gram grammar G. These WFSTs are composed and optimized into one H ? C ? L ? G WFST [8 … Cited by 1 – Related articles – All 4 versions

[PDF] Introduction to probabilistic models in Information retrieval [PDF] from ed.ac.uk V Lavrenko – proceedings of the 33rd International ACM …, 2010 – homepages.inf.ed.ac.uk Page 1. Copyright 2010, Victor Lavrenko Introduction to Probabilistic Models for Information Retrieval Victor Lavrenko University of Edinburgh Copyright 2010, Victor Lavrenko Objectives ? Highlight influential work on probabilistic models for IR … Cited by 1 – Related articles – View as HTML – All 3 versions

A comparison of audio-free speech recognition error prediction methods [PDF] from pitt.edu P Jyothi… – Tenth Annual Conference of the …, 2009 – isca-speech.org … Lm is a finite state automaton, and thus Lm -1 = Lm. The initial composition of W with Lm can be omitted without any trade-offs in accuracy if Lm is an n-gram grammar (as it is in our task) since it only serves to scale scores of W deterministi- cally. … Cited by 1 – Related articles – All 3 versions

System and methods for using transcripts to train an automated directory assistance service RM Schwartz, H Shu, J Makhoul… – US Patent 7,447,636, 2008 – Google Patents … The speech grammar estimation mod- ule (330) creates an n-gram grammar for multiple telephone numbers from the transcripts. … Page 4. TRANSCRIPTION 310 SPEECH ACOUSTIC MODEL TRAINING SPEECH GRAMMAR ESTIMATION 33Q N-GRAM GRAMMAR FIG. … Related articles – All 2 versions

[PDF] Speech recognition with weighted finite-state transducers [PDF] from ed.ac.uk M Mohri, F Pereira… – Handbook on Speech Processing …, 2008 – wiki.inf.ed.ac.uk … Our approach uses finite-state transducers, rather than acceptors, to represent the n-gram grammars, pronunciation dictionaries, context-dependency specifications, HMM topology, word, phone or HMM segmentations, lattices and n-best output lists encountered in ASR. … Cited by 40 – Related articles – View as HTML – All 8 versions

Training Statistical Language Models from Grammar-Generated Data: A Comparative Case-Study [PDF] from unige.ch B Hockey, M Rayner… – Advances in Natural Language …, 2008 – Springer … recognition always requires some kind of language model [1]. At least in the world of research, it has generally been assumed that language models are best constructed using some kind of data-driven process; the most common alternative in practice is the N-gram grammar. … Cited by 2 – Related articles – All 6 versions

VOICE PROCESSING DEVICE AND METHOD, AND PROGRAM K Minamino, H Honda, Y Maeda… – US Patent App. 12/ …, 2010 – Google Patents Page 1. (i9) United States (12) Patent Application Publication MINAMINO et al. US 20110029311A1 (io) Pub. No.: US 2011/0029311 Al (43) Pub. Date: Feb. 3, 2011 (54) VOICE PROCESSING DEVICE AND METHOD, AND PROGRAM … All 2 versions

Contextual filtering R Belvin, S Kadambe – US Patent 7,379,870, 2008 – Google Patents … Full syntactic parsing requires that a grammar be written for each appli- cation (and as noted, this is impracticable for dictation systems), and n-gram grammars are based on finding word- pair or word-triplet tokens in a training corpus. … Cited by 6 – Related articles – All 2 versions

Hippocampal-like categorization of object views: A self-organizing learning approach to vision modeling using stochastic grammar inference and associative memory PM Goebel, M Vincze… – … Technologies and Factory …, 2008 – ieeexplore.ieee.org … crossed. This yields in an elegant way the required transition matrix in order to define the stochastic n-gram grammar model. In the Figure 4, at node <W, Jn> two ways are possible, one branch pna and one branch pnb. Collecting … Related articles

Flexible and Expandable Speech Recognition Hardware with Weighted Finite State Transducers K You, J Choi… – Journal of Signal Processing Systems, 2011 – Springer … For example, four hierarchical knowledge sources commonly utilized in speech recog- nition can be represented by four separate WFSTs: HMM state structure H, context dependency C, pro- nunciation lexicon L, and n-gram grammar G. Then, these WFSTs are composed and … Related articles

Sign Language Recognition H Cooper, B Holt… – Visual Analysis of Humans, 2011 – Springer Page 1. Chapter 27 Sign Language Recognition Helen Cooper, Brian Holt, and Richard Bowden Abstract This chapter covers the key aspects of sign-language recognition (SLR), starting with a brief introduction to the motivations … Cited by 1 – Related articles

METHOD AND DATA PROCESSING SYSTEM FOR THE CONTROLLED QUERY OF STRUCTURED SAVED INFORMATION F Runge, F Burkhardt, J Liu… – US Patent App. 12/014,886, 2008 – Google Patents … from the ontology. [0060] Input grammar 320, which represents an input or speech model, may be designed, for example, as an EBNF grammar in text form and/or as a statistically based n-gram grammar. However, other grammar … All 2 versions

Method and apparatus for training an automated speech recognition-based system P Natarajan, R Prasad – US Patent 7,346,507, 2008 – Google Patents Page 1. Illllllllllllllllllllllllllllllllllllllllllllllllll US007346507B1 (12) United States Patent Natarajan et al. (io) Patent No.: (45) Date of Patent: US 7,346,507 Bl Mar. 18, 2008 (54) METHOD AND APPARATUS FOR TRAINING AN AUTOMATED … Cited by 2 – Related articles – All 2 versions

VEHICLE CONTROL RS Hernandez, D Scarlatti… – US Patent App. …, 2008 – freepatentsonline.com … There are several techniques for language modelling including grammar-based language modelling and statistical language modelling (N-gram). Grammar-based language modelling consists of defining all possible sentences that the system can recognise. … Cached

Rhythm and tempo analysis toward automatic music transcription [PDF] from korea.ac.kr H Takeda, T Nishimoto… – Acoustics, Speech and …, 2007 – ieeexplore.ieee.org … also works as meter analysis since bar line positions are then obtained as their boundaries, and time signature is obtained from the most likely rhythm words. The rhythm vocabulary and its N-gram grammar can be trained using original scores of composed music samples. … Cited by 4 – Related articles – All 4 versions

Robust location understanding in spoken dialog systems using intersections [PDF] from pitt.edu ML Seltzer, YC Ju, I Tashev… – Eighth Annual Conference …, 2007 – isca-speech.org … on the corner of <Street> and <Street>”. This intersec- tion CFG is then combined with a domain-independent n-gram filler model to generate a hybrid CFG/n-gram grammar which is much more robust than a standalone CFG to variations in the grammar [6]. … Cited by 4 – Related articles – All 8 versions

A framework for developing conversational user interfaces [PDF] from psu.edu J Glass, E Weinstein, S Cyphers, J Polifroni… – … -Aided Design of User …, 2005 – Springer … 3.1 Understanding The speech recogniser [11] uses generic telephone-based acoustic mod- els, phonetic descriptions of the words in the vocabulary, and an n-gram grammar, which provides likelihoods for sequences of n words, to describe the ways words occur in sentences. … Cited by 30 – Related articles – All 32 versions

[PDF] Combining linguistic and statistical methods for bi-directional English Chinese translation in the flight domain [PDF] from mit.edu S Seneff, C Wang… – … of the 7th Conference of the …, 2006 – people.csail.mit.edu … Multi-English Generation Rules Sentence Chinese Word Graph Class n-gram Grammar Chinese Grammar Interlingua English … Chinese Grammar Sentence Chinese Word Graph Class n-gram Grammar English English Translation Generation Rules English System II … Cited by 13 – Related articles – View as HTML – All 10 versions

LEARNING OF DIALOGUE STATES AND LANGUAGE MODEL OF SPOKEN INFORMATION SYSTEM DJ Attwater, MD Edgington… – EP Patent …, 2008 – freepatentsonline.com … call. Known techniques can then be used to generate a grammar, for example, a finite state network of dialogue states, or a bigram or n-gram grammar, for use in natural language automated call routing systems, for example. … Related articles – Cached

Lexical model for ASR of fluent Polish speech-analyze of polish phonology and phonetics K Rutecki… – Signals and Electronic Systems, 2008 …, 2008 – ieeexplore.ieee.org … A. System architecture model Most of state-of-the-art ASR is based on integrated archi- tecture [1,2]. This sort of ASR capture and compile into a single model several levels of linguistic structure. Know- ledge about words pronunciation is embedded into simple n-gram grammar. … Related articles

Unsupervised training in natural language call routing DJ McCarthy, P Natarajan – US Patent 7,092,888, 2006 – Google Patents … The best natural language call routing systems use statistical models both for speech recognition and language understanding. By using statistical n-gram grammars for speech recognition, it is possible to robustly model a huge set of possible customer queries. … Cited by 25 – Related articles – All 2 versions

Evaluation of modern-large vocabulary speech recognition techniques and their implementation [PDF] from sun.ac.za RA Swart – 2009 – ir1.sun.ac.za Page 1. Evaluation of modern large-vocabulary speech recognition techniques and their implementation by Renier Adriaan Swart Thesis presented in partial ful lment of the requirements for the degree of Master of Science in Electronic Engineering at … Related articles – All 4 versions

[PDF] A Weighted Finite State Transducer tutorial [PDF] from idiap.ch PN Garner – 2007 – idiap.ch … 6.2 Grammar In this example, almost the same ‘YES’ ‘NO’ lexicon as before is used, but this time with a completely hypothetical n-gram grammar: data ngram 1=5 ngram 2=6 … Note that this was generated by taking a real n-gram grammar and removing most of the words. … Cited by 1 – Related articles – View as HTML – All 10 versions

Understanding spoken location information based on intersections ML Seltzer, Y Ju… – US Patent App. 20,090/ …, 2007 – freepatentsonline.com … Street>”. This intersection CFG is then combined with a domain-independent N-gram filler model to generate a hybrid CFG/N-gram grammar which is much more robust than a stand alone CFG, to variations in the grammar. … Cached

A multi-pass, dynamic-vocabulary approach to real-time, large-vocabulary speech recognition [PDF] from pitt.edu IL Hetherington – Ninth European Conference on Speech …, 2005 – isca-speech.org … with a phone- mic bigram constraint. The n-gram grammar G is trained with OOV examples, allowing G to provide constraints on likely lo- cations of OOV occurrences during decoding. 2.4. Vocabulary Manipulation We have the … Cited by 12 – Related articles – All 12 versions

Pragmatics and computational linguistics [PDF] from psu.edu D Jurafsky – The handbook of pragmatics, 2004 – Wiley Online Library Page 1. 578 Daniel Jurafsky 26 Pragmatics and Computational Linguistics DANIEL JURAFSKY 1 Introduction These days there’s a computational version of everything. Computational biology, computational musicology, computational archaeology, and so on, ad infinitum. … Cited by 51 – Related articles – All 14 versions

Analysis of verbal and nonverbal acoustic signals with the Dresden UASR system [PDF] from pp.ua R Hoffmann, M Eichner… – Verbal and Nonverbal Communication …, 2007 – Springer … Recognizer Synthesizer Phrase level Language model (stochastic n-gram grammar and/or regular grammar) Rules or neural nets for determination of phrase boundaries and accents; language model in special cases, eg for concept-to-speech (CTS) … Cited by 9 – Related articles – BL Direct – All 3 versions

Natural Language Interface Control System CC Konopka… – US Patent App. 20,080/ …, 2007 – freepatentsonline.com … 10. The device of claim 8 wherein the speech recognition includes application of N gram grammars to the open-ended request. 11. … The speech recognition module 204 includes a speech decoder 216 , an N-gram grammar module 218 , and an acoustic models module 220 . … Cached

Natural Language Interface Control System CC KONOPKA… – US Patent App. 11/932,771, 2007 – Google Patents … loH: •us LZ S4 Feature Extraction N-gram Grammar ±1 Probabilistic Context Free Grammar Speech Decoder 1 11 T -2Vt Natural Language Interface Acoustic Models T Device Interface Feedback Module Controlled Device(s) >!l<f Device Abstraction(s) |^Ljp_I_| ||| Abstraction …

SPEECH ACT MODELING AND VERIFICATION IN SPOKEN DIALOGUE SYSTEMS CH Wu, JF Yeh… – Advances in Chinese spoken …, 2007 – books.google.com Page 339. CHAPTER 14 SPEECH ACT MODELING AND VERIFICATION IN SPOKEN DIALOGUE SYSTEMS Chung-Hsien Wu, Jui-Feng Yeh and Gwo-Lang Yan Department of Computer Science and Information Engineering, National Cheng Rung University, No. … Related articles

Establishing a multimodal application voice CW Cross, MC Hollinger, IR Jablokov… – US Patent App. 11/ …, 2005 – Google Patents Page 1. US 20060287865A1 (i9) United States (12) Patent Application Publication oo) Pub. No.: US 2006/0287865 Al Cross, JR. et al. (43) Pub. Date: Dec. 21,2006 (54) ESTABLISHING A MULTIMODAL APPLICATION VOICE … Cited by 13 – Related articles – All 2 versions

Methods and Systems for Query-Based Searching Using Spoken Input IM Bennett – US Patent App. 20,100/228,540, 2010 – freepatentsonline.com Systems and methods for query-based searching using spoken input are disclosed. In systems and methods according to embodiments of the invention, continuous speech natural language queries are accepte. Cached

Analysis of Verbal and Nonverbal Acoustic Signals with the Dresden UASR System R Ho?mann, M Eichner… – COST Action, 2007 – books.google.com … The basic idea which we published ?rstly in [2] is illustrated in Figure 1. Table 1. Selected databases in speech recognizers and synthesis Recognizer Synthesizer Language model (stochastic n-gram grammar and/or regular grammar) Rules or neural nets for determination of … Cited by 2 – Related articles

Embedded speech recognition systems [PDF] from auckland.ac.nz O Cheng – 2008 – researchspace.auckland.ac.nz Page 1. http://researchspace.auckland.ac.nz ResearchSpace@Auckland Copyright Statement The digital copy of this thesis is protected by the Copyright Act 1994 (New Zealand). This thesis may be consulted by you, provided … Cited by 1 – Related articles – Library Search – All 4 versions

Methods and Systems for Searching Using Spoken Input and User Context Information IM Bennett – US Patent App. 20,100/235,341, 2010 – freepatentsonline.com Systems and methods for searching using spoken input and user context information are disclosed. In systems and methods according to embodiments of the invention, continuous speech natural language qu. Cached

[PDF] A virtual butler controlled by speech [PDF] from unizar.es A Uria, A Ortega, MI Torres, A Miguel… – Proceedings of the III …, 2006 – jth2006.unizar.es … authors have contributed to the development of these models [1, 2, 3]. The second level is the language model and referes to how words are concatenated, which can as simple as static rules of grammar, or more complex as an statistically es- timated N-gram grammar [4]. This … Cited by 2 – Related articles – View as HTML – All 5 versions

A framework for predicting speech recognition errors E Fosler-Lussier, I Amdal… – Speech communication, 2005 – Elsevier … canonical phone. In this case, Eq. (2) can be simplified (also eliminating the initial deterministic scaling by the n-gram grammar L): (3)W conf =W ring operator P -1 ring operator C ring operator P ring operator L. Besides the inclusion … Cited by 11 – Related articles – All 2 versions

Some approaches to statistical and finite-state speech-to-speech translation F Casacuberta, H Ney, FJ Och, E Vidal… – Computer Speech & …, 2004 – Elsevier Cited by 72 – Related articles – All 3 versions

Generation of word alternative pronunciations using weighted finite state transducers [PDF] from pitt.edu S Paulo… – Ninth European Conference on Speech …, 2005 – isca-speech.org … Such a system can perform poorly if higher level and more reli- able information is not used. In [4], a phone recognizer is used together with an n-gram grammar of phones to achieve reliable corpus transcriptions based on the signal acoustics with an it- erative algorithm. … Cited by 10 – Related articles – All 7 versions

High-performance tagging on medical texts [PDF] from upenn.edu U Hahn… – Proceedings of the 20th international conference …, 2004 – dl.acm.org … in MED. Another indicator of a simpler POS n-gram grammar in medical nar- ratives is the fact that the absolute number of POS n-gram types common to NEGRA and MED is much lower than for NEGRA and NEWS. Scaled against … Cited by 11 – Related articles – All 19 versions

Learning of dialogue states and language model of spoken information system DJ Attwater, MD Edgington, PJ Durston – US Patent 6,839,671, 2005 – Google Patents Page 1. Illllllllllllllllllllllllllllllllllllllllllllllllll US006839671B2 (12) United States Patent ao) Patent No.: us 6,839,671 B2 Attwater et al. (45) Date of Patent: Jan. 4,2005 (54) LEARNING OF DIALOGUE STATES AND LANGUAGE MODEL … Cited by 17 – Related articles – All 4 versions

Tagging medical documents with high accuracy U Hahn… – PRICAI 2004: Trends in Artificial Intelligence, 2004 – Springer … also favors high-occurring (more than 1000 times) types in MED. Another indicator of a simpler POS n-gram grammar in medical narratives is the fact that the absolute num- ber of POS n-gram types common to NEGRA and MED is much lower than for NEGRA and NEWS. … Cited by 6 – Related articles – BL Direct – All 2 versions

Robust access to large structured data using voice form-filling [PDF] from pitt.edu S Parthasarathy, C Allauzen… – … European Conference on …, 2005 – isca-speech.org … A task specific phonotactic grammar (statis- tical N-gram grammars) could be used to output a phone lattice in the first-pass, and both lexical and grammatical constraints could be incorporated by rescoring in a second-pass. … Cited by 3 – Related articles – All 5 versions

Personalized Computer Architecture as Contextual Partitioning for Speech Recognition [PDF] from vt.edu CG Kent – 2009 – scholar.lib.vt.edu … Grammar models constrain the recognition process, and play a significant role in achieving high system accuracy. With contextual partitioning however, a N-gram grammar model would not be solely sufficient. Particularly with fine-grained CP, there could be significant interplay … Related articles – Library Search

Open Microphone Speech Understanding: Correct Discrimination of in Domain Speech J Hieronymus, J Dowding – Acoustics, Speech and Signal …, 2006 – ieeexplore.ieee.org … A class N-gram grammar has been trained on the same training data. … 2. RULE BASED LANGUAGE MODELS Most of the present dialogue systems use either n-gram grammars or hand built finite state grammars for speech recognition. … Related articles

[PDF] The Juicer LVCSR decoder-user manual [PDF] from epfl.ch D Moore – IDIAP Research Institute, Martigny, Switzerland, 2005 – infoscience.epfl.ch Page 1. I C O M M U N C A T I O N I D I A P Rue du Simplon 4 IDIAP Research Institute 1920 Martigny – Switzerland www.idiap.ch Tel: +41 27 721 77 11 Email: info@idiap.ch PO Box 592 Fax: +41 27 721 77 12 The Juicer LVCSR Decoder – User Manual for Juicer version 0.5.0 … Cited by 3 – Related articles – View as HTML – All 7 versions

Semantic Decoding of User Queries IM Bennett – US Patent App. 20,090/157,401, 2008 – freepatentsonline.com An intelligent query system for processing voiced-based queries is disclosed, which uses semantic based processing to identify the question posed by the user by understanding the meaning of the user’s. Cached

Browsing and Retrieval of Full Broadcast-Quality Video A Basso, MR Civanlar, DC Gibbon… – US Patent App. 12/ …, 2010 – Google Patents … Granada), Spain, 1993, incorporated herein by reference. Stochastic n-gram grammars represent portions of the sentence (ie, concepts) that have associations to defined conceptual units. Some of them are task independent … All 3 versions

A hybrid dialogue strategy for speech-enabled mobile commerce Y Fan… – e-Business Engineering, 2005. ICEBE …, 2005 – ieeexplore.ieee.org … The system uses N-Gram grammars to govern the recognition at the request segment of a dialogue, and employs an application-directed strategy at the clarification discourse segment. The paper also details generating a corpus … Related articles – All 3 versions

A fast HMM match algorithm for very large vocabulary speech recognition A Seward – Speech communication, 2004 – Elsevier … This is especially apparent in tasks that incorporate large vocabularies and long-dependency n-gram grammars, since these impose a high degree of context dependency and HMMs have to be treated differently in each context. … Cited by 5 – Related articles – All 4 versions

[PDF] Building a visual speech recognizer [PDF] from tudelft.nl KF Driel – Delft University of Technology, 2009 – repository.tudelft.nl Page 1. Building a visual speech recognizer Master’s thesis Karin F. Driel August 18, 2009 Page 2. Page 3. 3 Building a visual speech recognizer Thesis, submitted in partial fulfillment of the requirements for the degree of Master of Science in Media & Knowledge Engineering … Cited by 1 – Related articles – View as HTML – All 3 versions

Statistical Language Model Trained With Semantic Variants IM Bennett – US Patent App. 20,080/052,078, 2007 – freepatentsonline.com An intelligent query system for processing voiced-based queries is disclosed, which uses a combination of both statistical and semantic based processing to identify the question posed by the user by u. Cached

Browsing and retrieval of full broadcast-quality video A Basso, MR Civanlar, DC Gibbon… – US Patent …, 2011 – Google Patents Page 1. US007877774B1 (12) United States Patent ao) Patent no.: us 7,877,774 bi Basso etal. (45) Date of Patent: Jan. 25,2011 (54) BROWSING AND RETRIEVAL OF FULL BROADCAST-QUALITY VIDEO (75) Inventors: Andrea … Cited by 1 – Related articles – All 2 versions

Natural Language Speech Lattice Containing Semantic Variants IM Bennett – US Patent App. 20,080/059,153, 2007 – freepatentsonline.com An intelligent query system for processing voiced-based queries is disclosed, which uses a combination of both statistical and semantic based processing to identify the question posed by the user by u. Cached

[PDF] Speech enhancement using microphone array [PDF] from ohiolink.edu J Cho – 2005 – etd.ohiolink.edu Related articles – All 4 versions

Detail publikace P Aleš, I Pavel, Š Jan – 2008 – kky.zcu.cz … Nakladatel: IEEE. Místo vydání: Peking, Cína. Datum: 29.10.2008. 2011-03-15 16:21:32 / 2011-06-09 12:30:02 / 1. Klícová slova. language modeling, n-gram, grammar, real-time, LVCSR. Klícová slova v ceštine. jazykové modelování, n-gram, gramatika, reálný cas, LVCSR. BibTeX … Cached

[PDF] The construction and evaluation of statistical models of melodic structure in music perception and composition [PDF] from gold.ac.uk MT Pearce – City University, London, 2005 – doc.gold.ac.uk Page 1. THE CONSTRUCTION AND EVALUATION OF STATISTICAL MODELS OF MELODIC STRUCTURE IN MUSIC PERCEPTION AND COMPOSITION Marcus Thomas Pearce Doctor of Philosophy Department of Computing City University, London December 2005 Page 2. … Cited by 26 – Related articles – View as HTML – All 3 versions

System and method for a spoken language interface to a large database of changing records Y Lyudovyk – US Patent App. 10/840,377, 2004 – Google Patents … The recognition of the user’s input may be carried out using a grammar database 120. [0028] As an example, the grammar database 120 may be a statistical N-gram grammar such as a uni-gram grammar, bi-gram grammar, tri-gram grammar, etc. …

System and method for a spoken language interface to a large database of changing records Y Lyudovyk – US Patent App. 20,050/004,799, 2004 – freepatentsonline.com … The recognition of the user’s input may be carried out using a grammar database 120 . As an example, the grammar database 120 may be a statistical N-gram grammar such as a uni-gram grammar, bi-gram grammar, tri-gram grammar, etc. … Cached

[CITATION] Language Modeling in Information Retrieval V Januzaj – 2004 Related articles – All 2 versions

Towards the Development of an Automatic Diacritizer for the Persian Orthography based on the Xerox Finite State Transducer [PDF] from uottawa.ca P Nojoumian – 2011 – ruor.uottawa.ca Page 1. Towards the Development of an Automatic Diacritizer for the Persian Orthography based on the Xerox Finite State Transducer Thesis submitted to the Faculty of Graduate and Postdoctoral Studies In partial fulfillment of the requirements For the PhD degree in Linguistics … All 2 versions

Method for processing speech using dynamic grammars IM Bennett – US Patent App. 20,040/236,580, 2004 – freepatentsonline.com Speech data is processed with one or more dynamic grammars, to reduce latency and improve accuracy. Different speech grammars are used by a speech recognition process depending on a context experience. Cached

[PDF] Cross-document coreference between different types of collateral texts for films [PDF] from open.ac.uk E Tomadaki – PhD diss., University of Surrey, 2006 – people.kmi.open.ac.uk Page 1. Cross-Document Coreference between Different Types of Collateral Texts for Films Eleftheria Tomadaki Submitted to the University of Surrey for the Degree of Doctor of Philosophy Department of Computing School of … Cited by 4 – Related articles – View as HTML – All 3 versions

???????????????????? N-gram ? Grammar ??? ???? – 2006 – library.naist.jp … ?????: Combining grammar with N-gram for enhancing a public speech dialogue system. ??: ??, ??????, ????. ?????: ?????????????????????? N-gram Grammar. ???: 24-Mar-2006. ???: ?????????????. … Cached

Network based interactive speech recognition system IM Bennett – US Patent App. 20,060/235,696, 2006 – freepatentsonline.com A network based interactive speech system responds in real-time to speech-based queries addressed to a set of topic entries. A best matching response is provided based on speech recognition processing. Cached

[BOOK] Speech processing for IP networks: Media resource control protocol (MRCP) D Burke – 2007 – books.google.com Page 1. Speech Processing for IP Networks Media Resource Control Protocol (MRCP) Dave Burke Page 2. Page 3. Speech Processing for IP Networks Page 4. Page 5. Speech Processing for IP Networks Media Resource Control Protocol (MRCP) Dave Burke Page 6. … Cited by 7 – Related articles – Library Search – All 6 versions

[PDF] Audient: An Acoustic Search Engine [PDF] from tedleath.com T Leath – 2005 – tedleath.com … The term “stochastic” suggests estimation of probability and/or guesswork. An N-Gram grammar can model the probability of a word based on the prior occurrence of N-1 other words. Unigrams examine only the current word while bigrams rely on the likelihood of word pairs, 8 … Cited by 2 – Related articles – View as HTML – All 7 versions

[PDF] Identifying Languages in the Limit Using Alignment-based Learning [PDF] from psu.edu BCSB Eng – 2004 – Citeseer … referring to a stochastic bigram grammar. Bigram grammars are a special case of N-gram grammars, where the N is related to the number of previous symbols that affect the probability of the next symbol. For any two symbols ? … Cited by 3 – Related articles – View as HTML – Library Search – All 4 versions

System & method for processing sentence based queries IM Bennett – US Patent App. 20,050/086,049, 2004 – freepatentsonline.com Sentence based queries from a user are analyzed to determine appropriate answers from an electronic database. Text in the queries is dissected to identify different phrases; the phrases in turn are ma. Cached

Method for processing speech signal features for streaming transport IM Bennett – US Patent App. 20,040/249,635, 2004 – freepatentsonline.com Speech signal information is formatted, processed and transported in accordance with a format adapted for TCP/IP protocols used on the Internet and other communications networks. NULL characters are u. Cached

System & method for natural language processing of sentence based queries IM Bennett – US Patent App. 20,050/086,046, 2004 – freepatentsonline.com Sentence based queries from a user are analyzed using a natural language engine to determine appropriate answers from an electronic database. The system and methods are useful for Internet based searc. Cached

System & method for natural language processing of query answers IM Bennett – US Patent App. 20,050/080,614, 2004 – freepatentsonline.com Candidate answers responsive to a user query are analyzed using a natural language engine to determine appropriate answers from an electronic database. The system and methods are useful for Internet b. Cached

Partial speech processing device & method for use in distributed systems IM Bennett – US Patent App. 20,050/086,059, 2004 – freepatentsonline.com A client device incorporates partial speech recognition for recognizing a spoken query by a user. The full recognition process is distributed over a client/server architecture, so that the amount of p. Cached

[PDF] Dialectal chinese speech recognition: Final report [PDF] from ogi.edu R Sproat, F Zheng, L Gu, D Jurafsky… – … , Baltimore, MD, Tech. …, 2004 – csee.ogi.edu … native English. Next, the 6 phone recognizers were run on the 2 training sets, generating 6 sets of phone output and 6 sets of phone N-gram grammars, one set for native English and one set for Japanese-accented En- glish. In … Cited by 6 – Related articles – View as HTML – All 7 versions

Speech, Text and Braille Conversion Technology R Hoffmann – Assistive Technology for Visually Impaired and Blind …, 2008 – Springer … types of systems. Table 14.2. Databases for speech recognizers and synthesizers Recognizer Synthesizer Sentence level Language model Stochastic n-gram grammar Regular (automata) grammar Rules for detecting phrase … Related articles – Library Search

Building a Standards and Research Community with the Galaxy Communicator Software Infrastructure S Bayer – Practical Spoken Dialog Systems, 2004 – Springer Page 1. Chapter 9 Building a Standards and Research Community with the Galaxy Communicator Software Infrastructure Samuel Bayer The Mitre Corporation Key words: dialog, standards, software infrastructure Abstract: The … Cited by 1 – Related articles – All 3 versions