Perplexity—a measure of the difficulty of speech recognition tasks

Jelinek, F.; Mercer, R.L.; Bahl, Lalit R.; Baker, James K.

doi:10.1121/1.2016299

Cited by 226 publications

(112 citation statements)

References 0 publications

Supporting

Mentioning

110

Contrasting

Unclassified

Order By: Relevance

“…5) containing each trigram at least once. The perplexity, 2 (where is entropy [22]), of this vocabulary is 5.53 per digram. By comparison, the perplexity of an equivalent model built from the 1,000 most common English words is 10.31.…”

Section: Resultsmentioning

confidence: 99%

Fingerspelling Recognition through Classification of Letter-to-Letter Transitions

Ricco

Tomasi

2010

Computer Vision – ACCV 2009

View full text Add to dashboard Cite

Abstract. We propose a new principle for recognizing fingerspelling sequences from American Sign Language (ASL). Instead of training a system to recognize the static posture for each letter from an isolated frame, we recognize the dynamic gestures corresponding to transitions between letters. This eliminates the need for an explicit temporal segmentation step, which we show is error-prone at speeds used by native signers. We present results from our system recognizing 82 different words signed by a single signer, using more than an hour of training and test video. We demonstrate that recognizing letter-to-letter transitions without temporal segmentation is feasible and results in improved performance.

show abstract

Section: Resultsmentioning

confidence: 99%

Fingerspelling Recognition through Classification of Letter-to-Letter Transitions

Ricco

Tomasi

2010

Computer Vision – ACCV 2009

View full text Add to dashboard Cite

show abstract

“…, w m ) À1/m , is to maximize probability, because PPL is the inverse probability of the test set, normalized by the number of words. That is to say, a lower PPL indicates a better model [34]. If a language model built from the augmented corpus shows improved perplexity for the test set, it indicates the usefulness of our approach for corpus expansion.…”

Section: Discussionmentioning

confidence: 99%

Sentence-Chain Based Seq2seq Model for Corpus Expansion

Chung¹,

Park²

2017

ETRI Journal

View full text Add to dashboard Cite

This study focuses on a method for sequential data augmentation in order to alleviate data sparseness problems. Specifically, we present corpus expansion techniques for enhancing the coverage of a language model. Recent recurrent neural network studies show that a seq2seq model can be applied for addressing language generation issues; it has the ability to generate new sentences from given input sentences. We present a method of corpus expansion using a sentence-chain based seq2seq model. For training the seq2seq model, sentence chains are used as triples. The first two sentences in a triple are used for the encoder of the seq2seq model, while the last sentence becomes a target sequence for the decoder. Using only internal resources, evaluation results show an improvement of approximately 7.6% relative perplexity over a baseline language model of Korean text. Additionally, from a comparison with a previous study, the sentence chain approach reduces the size of the training data by 38.4% while generating 1.4-times the number of ngrams with superior performance for English text.

show abstract

“…One of the most widely used speech recognition difficulty measures at present is the so-called perplexity, which was introduced by Jelinek in 1977 [5]. Let W be a word sequence, w1; w2 ; 1 1 1 ; wn; allowed by the grammar; perplexity is then defined as P P = P (w1; 1 1 1 ; wn)…”

Section: Difficulty Measures For Speech Recognitionmentioning

confidence: 99%

A transcription-based approach to determine the difficulty of a speech recognition task

Garcı́a

Rubio²,

Díaz-Verdejo³

et al. 1999

IEEE Trans. Speech Audio Process.

View full text Add to dashboard Cite

Abstract-A new parameter for estimating the difficulty of a continuous speech recognition task, called speech decoding difficulty, is presented in this work. It is obtained from the language model defined for the recognition task and the phonetic similarity between the transcriptions of the words that make up the vocabulary used. Two variants of the proposed task difficulty measure are introduced: ideal speech decoding difficulty (ISDD), which is uninfluenced by practical considerations on the recognition system implemented, and a second, more realistic variant, called practical speech decoding difficulty (PSDD) to study the performance of a specific recognition system when confronting a given task.

show abstract

Perplexity—a measure of the difficulty of speech recognition tasks

Cited by 226 publications

References 0 publications

Fingerspelling Recognition through Classification of Letter-to-Letter Transitions

Fingerspelling Recognition through Classification of Letter-to-Letter Transitions

Sentence-Chain Based Seq2seq Model for Corpus Expansion

A transcription-based approach to determine the difficulty of a speech recognition task

Contact Info

Product

Resources

About