The problem of estimating the parameter values of hidden Markov word models for speech recognition is addressed. It is argued that maximum-likelihood estimation of the parameters via the forward-backward algorithm may not lead to values which maximize recognition accuracy. An alternative estimation procedure called corrective training, which is aimed at minimizing the number of recognition errors, is described. Corrective training is similar to a well-known error-correcting training procedure for linear classifiers and works by iteratively adjusting the parameter values so as to make correct words more probable and incorrect words less probable. There are strong parallels between corrective training and maximum mutual information estimation; the relationship of these two techniques is discussed and a comparison is made of their performance. Although it has not been proved that the corrective training algorithm converges, experimental evidence suggests that it does, and that it leads to fewer recognition errors that can be obtained with conventional training methods.>

A technique for constructing Markov models for the acoustic representation of words is described. Word models are constructed from models of subword units called fenones. Fenones represent very short speech events and are obtained automatically through the use of a vector quantizer. The fenonic baseform for a word-i.e., the sequence of fenones used to represent the word-is derived automatically from one or more utterances of that word. Since the word models are all composed from a small inventory of subword models, training for large-vocabulary speech recognition systems can be accomplished with a small training script. A method for combining phonetic and fenonic models is presented. Results of experiments with speaker-dependent and speaker-independent models on several isolated-word recognition tasks are reported. The results are compared with those for phonetics-based Markov models and template-based dynamic programming (DP) matching.>

In this paper we describe a statistical technique for aligning sentences with their translations in two parallel corpora.In addition to certain anchor points that are available in our da.ta, the only information about the sentences that we use for calculating alignments is the number of tokens that they contain.Because we make no use of the lexical details of the sentence, the alignment computation is fast and therefore practical for application to very large collections of text.We have used this technique to align several million sentences in the English-French Hans~trd corpora and have achieved an accuracy in excess of 99% in a random selected set of 1000 sentence pairs that we checked by hand.We show that even without the benefit of anchor points the correlation between the lengths of aligned sentences is strong enough that we should expect to achieve an accuracy of between 96% and 97%.Thus, the technique may be applicable to a wider variety of texts than we have yet tried.

We describe a statistical technique for assigning senses to words. An instance of a word is assigned a sense by asking a question about the context in which the word appears. The question is constructed to have high mutual information with the translation of that instance in another language. When we incorporated this method of assigning senses into our statistical machine translation system, the error rate of the system decreased by thirteen percent.

A description is presented of the authors' current research on automatic speech recognition of continuously read sentences from a naturally-occurring corpus: office correspondence. The recognition system combines features from their current isolated-word recognition system and from their previously developed continuous-speech recognition system. It consists of an acoustic processor, an acoustic channel model, a language model, and a linguistic decoder. Some new features in the recognizer relative to the isolated-word speech recognition system include the use of a fast match to prune rapidly to a manageable number the candidates considered by the detailed match, multiple pronunciations of all function words, and modeling of interphone coarticulatory behavior. The authors recorded training and test data from a set of ten male talkers. The perplexity of the test sentences was found to be 93; none of sentences was part of the data used to generate the language model. Preliminary (speaker-dependent) recognition results on these talkers yielded an average word error rate of 11.0%.>

The acoustic-modelling problem in automatic speech recognition is examined from an information theoretic point of view. This problem is to design a speech-recognition system which can extract from the speech waveform as much information as possible about the corresponding word sequence. The information extraction process is factored into two steps: a signal-processing step which converts a speech waveform into a sequence of informative acoustic feature vectors, and a step which models such a sequence. The authors are primarily concerned with the use of hidden Markov models to model sequences of feature vectors which lie in a continuous space. They explore the trade-off between packing information into such sequences and being able to model them accurately. The difficulty of developing accurate models of continuous-parameter sequences is addressed by investigating a method of parameter estimation which is designed to cope with inaccurate modeling assumptions.>

Discusses the problem of estimating the parameter values of hidden Markov word models for speech recognition. The authors argue that maximum-likelihood estimation of the parameters does not lead to values which maximize recognition accuracy and describe an alternative estimation procedure called corrective training which is aimed at minimizing the number of recognition errors. Corrective training is similar to a well-known error-correcting training procedure for linear classifiers and works by iteratively adjusting the parameter values so as to make correct words more probable and incorrect words less probable. There are also strong parallels between corrective training and maximum mutual information estimation. They do not prove that the corrective training algorithm converges, but experimental evidence suggests that it does, and that it leads to significantly fewer recognition errors than maximum likelihood estimation.>

The Speech Recognition Group at IBM Research has developed a real-time, isolated-word speech recognizer called Tangora, which accepts natural English sentences drawn from a vocabulary of 20000 words. Despite its large vocabulary, the Tangora recognizer requires only about 20 minutes of speech from each new user for training purposes. The accuracy of the system and its ease of training are largely attributable to the use of hidden Markov models in its acoustic match component. An automatic technique for constructing Markov word models is described and results are included of experiments with speaker-dependent and speaker-independent models on several isolated-word recognition tasks.>

The Speech Recognition Group at IBM Research in Yorktown Heights has developed a real-time, isolated-utterance speech recognizer for natural language based on the IBM Personal Computer AT and IBM Signal Processors. The system has recently been enhanced by expanding the vocabulary from 5,000 words to 20,000 words and by the addition of a speech workstation to support usability studies on document creation by voice. The system supports spelling and interactive personalization to augment the vocabularies. This paper describes the implementation, user interface, and comparative performance of the recognizer.

The Speech Recognition Group at IBM Research in Yorktown Heights has designed a real-time, isolated-utterance speech recognizer for natural language with a 5,000-word vocabulary based on the IBM Personal Computer (PC) AT model and two IBM Signal Processors realized in VLSI technology. The enrollment period for a new user is approximately 20 minutes. The basic vocabulary is chosen from the most common words in several collections of documents such as office memoranda and business letters. The system supports spelling and interactive personalization to augment this vocabulary. Signal processing, vector quantization, and acoustic matching algorithms are programmed on the IBM Signal Processors which fit into the PC AT chassis. The PC AT controls the Processors and implements the decoder stack search and the language model, as well as the application-specific interface. The modular architecture of the design is expandable to a 20,000-word vocabulary system by the addition of two more IBM Signal Processors housed in a PC Expansion Unit.

A method for estimating the parameters of hidden Markov models of speech is described. Parameter values are chosen to maximize the mutual information between an acoustic observation sequence and the corresponding word sequence. Recognition results are presented comparing this method with maximum likelihood estimation.

In order to achieve state-of-the-art performance in a speaker-dependent speech recognition task, it is necessary to collect a large number of acoustic data samples during the training process. Providing these samples to the system can be a long and tedious process for users. One way to attack this problem is to make use of extra information from a data bank representing a large population of speakers. In this paper we demonstrate that by using Bayesian techniques, prior knowledge derived from speaker-independent data can be combined with speaker-dependent training data to improve system performance.

Dynamic programming is used in speech recognition to search efficiently for word sequences whose templates best match acoustic data. The search is constrained by finite-state networks embodying grammatical rules. Typically, dynamic programming is implemented in two steps: the first calculates, for each state in a network and for each time, the best way of arriving at that state at that time; the second traces back from the final state at the final time to the initial state at the initial time to determine the best path through the network. This second step cannot be initiated before the determination (usually from the detection of silence) that the final state has been reached. Such a determination is difficult in the recognition of truly continuous speech; there are often no reliable anchor points. Further, it is often desirable to be able to recognize at least part of an utterance before a speaker has stopped talking. In this paper we introduce a technique for discovering the initial section of the optimal path through a network before the traversal of the network is complete. It can be used to report a system's interpretation of acoustic data from the not-too-distant past without relying on or making any decisions which may degrade recognition accuracy.

To evaluate the performance of a speech recognition system, large databases of labeled speech, including various speakers, noise conditions, and vocabularies, are necessary. This paper describes a method for automatically labeling speech data. In the past, speech has been labeled manually, typically by listening to and viewing waveforms through real-time, interactive computer I/O stations. This process is slow and tedious, and accounts for the shortage of large speech databases. The automatic labeling method reported here uses dynamic programming to align a script which is produced as the output of a recognition system, and a known script. The alignment gives a tentative labeling which can be refined by repeating the training, recognition, and alignment processes. The method was used to label a 50 speaker database of 140,000 digits.