Back-off Language Models Research Articles

A Comparison of Phrase Based and Word based Language Model for Punjabi

A language model provides connection to the decoding process to determine a precise word from several available options in the information base or phrase table. The language model can be generated using n-gram approach. Various language models and smoothing procedures are there to determine this model, like unigram, bigram, trigram, interpolation, backoff language model etc. We have done some experiments with different language models where we have used phrases in place of words as the smallest unit. Experiments have shown that phrase based language model yield more accurate results as compared to simple word based mode. We have also done some experiments with machine translation system where we have used phrase based language model rather than word based model and system yield great improvement.

Getting more from automatic transcripts for semi-supervised language modeling

Many under-resourced languages such as Arabic diglossia or Hindi sub-dialects do not have sufficient in-domain text to build strong language models for use with automatic speech recognition (ASR). Semi-supervised language modeling uses a speech-to-text system to produce automatic transcripts from a large amount of in-domain audio typically to augment a small amount of manual transcripts. In contrast to the success of semi-supervised acoustic modeling, conventional language modeling techniques have provided only modest gains. This paper first explains the limitations of back-off language models due to their dependence on long-span n-grams, which are difficult to accurately estimate from automatic transcripts. From this analysis, we motivate a more robust use of the automatic counts as a prior over the estimated parameters of a log-linear language model. We demonstrate consistent gains for semi-supervised language models across a range of low-resource conditions.

Converting Neural Network Language Models into Back-off Language Models for Efficient Decoding in Automatic Speech Recognition

Neural network language models (NNLMs) have achieved very good performance in large-vocabulary continuous speech recognition (LVCSR) systems. Because decoding with NNLMs is computationally expensive, there is interest in developing methods to approximate NNLMs with simpler language models that are suitable for fast decoding. In this work, we propose an approximate method for converting a feedforward NNLM into a back-off n-gram language model that can be used directly in existing LVCSR decoders. We convert NNLMs of increasing order to pruned back-off language models, using lower-order models to constrain the n-grams allowed in higher-order models. In experiments on Broadcast News data, we find that the resulting back-off models retain the bulk of the gain achieved by NNLMs over conventional n-gram language models, and give accuracy improvements as compared to existing methods for converting NNLMs to back-off models. In addition, the proposed approach can be applied to any type of non-back-off language model to enable efficient decoding.

Error handling approach using characterization and correction steps for handwritten document analysis

In this paper, we present a framework to handle recognition errors from a N-best list of output phrases given by a handwriting recognition system, with the aim to use the resulting phrases as inputs to a higher-level application. The framework can be decomposed into four main steps: phrase alignment, detection, characterization, and correction of word error hypotheses. First, the N-best phrases are aligned to the top-list phrase, and word posterior probabilities are computed and used as confidence indices to detect word error hypotheses on this top-list phrase (in comparison with a learned threshold). Then, the errors are characterized into predefined types, using the word posterior probabilities of the top-list phrase and other features to feed a trained SVM. Finally, the final output phrase is retrieved, thanks to a correction step that used the characterized error hypotheses and a designed word-to-class backoff language model. First experiments were conducted on the ImadocSen-OnDB handwritten sentence database and on the IAM-OnDB handwritten text database, using two recognizers. We present first results on an implementation of the proposed framework for handling recognition errors on transcripts of handwritten phrases provided by recognition systems.

Continuous space language models

This paper describes the use of a neural network language model for large vocabulary continuous speech recognition. The underlying idea of this approach is to attack the data sparseness problem by performing the language model probability estimation in a continuous space. Highly efficient learning algorithms are described that enable the use of training corpora of several hundred million words. It is also shown that this approach can be incorporated into a large vocabulary continuous speech recognizer using a lattice rescoring framework at a very low additional processing time. The neural network language model was thoroughly evaluated in a state-of-the-art large vocabulary continuous speech recognizer for several international benchmark tasks, in particular the N ist evaluations on broadcast news and conversational speech recognition. The new approach is compared to four-gram back-off language models trained with modified Kneser–Ney smoothing which has often been reported to be the best known smoothing method. Usually the neural network language model is interpolated with the back-off language model. In that way, consistent word error rate reductions for all considered tasks and languages were achieved, ranging from 0.4% to almost 1% absolute.

Improved modeling and efficiency for automatic transcription of Broadcast News

Over the last few years, the DARPA-sponsored Hub-4 continuous speech recognition evaluations have advanced speech recognition technology for automatic transcription of broadcast news. In this paper, we report on our research and progress in this domain, with an emphasis on efficient modeling with significantly fewer parameters for faster and more accurate recognition. In the acoustic modeling area, this was achieved through new parameter tying, Gaussian clustering, and mixture weight thresholding schemes. The effectiveness of acoustic adaptation is greatly increased through unsupervised clustering of test data. In language modeling, we explored the use of non-broadcast-news training data as well as the adaptation to topic and speaking styles. We developed an effective and efficient parameter pruning technique for backoff language models that allowed us to cope with ever increasing amounts of training data and expanded N-gram scopes. Finally, we improved our progressive search architecture with more efficient algorithms for lattice generation, compaction, and incorporation of higher-order language models.

Similarity-Based Models of Word Cooccurrence Probabilities

In many applications of natural language processing (NLP) it is necessary to determine the likelihood of a given word combination. For example, a speech recognizer may need to determine which of the two word combinations “eat a peach” and ”eat a beach” is more likely. Statistical NLP methods determine the likelihood of a word combination from its frequency in a training corpus. However, the nature of language is such that many word combinations are infrequent and do not occur in any given corpus. In this work we propose a method for estimating the probability of such previously unseen word combinations using available information on “most similar” words. We describe probabilistic word association models based on distributional word similarity, and apply them to two tasks, language modeling and pseudo-word disambiguation. In the language modeling task, a similarity-based model is used to improve probability estimates for unseen bigrams in a back-off language model. The similarity-based method yields a 20% perplexity improvement in the prediction of unseen bigrams and statistically significant reductions in speech-recognition error. We also compare four similarity-based estimation methods against back-off and maximum-likelihood estimation methods on a pseudo-word sense disambiguation task in which we controlled for both unigram and bigram frequency to avoid giving too much weight to easy-to-disambiguate high-frequency configurations. The similarity-based methods perform up to 40% better on this particular task.