Before next-generation human language technology can be designed to function successfully in actual #eld settings, interface techniques will be needed that can guide users' language to coincide with current system capabilities. The present study examines how input modality and presentation structure in#uence the linguistic complexity observed in people's spoken and written input to an interactive system. Using a semi-automatic simulation technique, language was collected during speech-only, writing-only, and combined pen#voice exchanges, and using presentation formats that either were structured or unconstrained. Results indicate that both modality and presentation format substantially in#uence linguistic complexity, although the speci#c nature of their impact di#ers. A comprehensive analysis is provided of how both factors a#ect people's observed language in terms of total words, dis#uencies, utterance length, lexical variability, perplexity, syntactic ambiguity, and semanti...

. Machine learning can be a most valuable tool for improvingthe flexibility and efficiency of robot applications. Many approaches to applying machine learning to robotics are known. Some approaches enhance the robot's high-level processing, the planning capabilities. Other approaches enhance the low-level processing, the control of basic actions. In contrast, the approach presented in this paper uses machine learning for enhancing the link between the low-level representations of sensing and action and the high-level representation of planning. The aim is to facilitate the communication between the robot and the human user. A hierarchy of concepts is learned from route records of a mobile robot. Perception and action are combined at every level, i.e., the concepts are perceptually anchored. The relational learning algorithm grdt has been developed which completely searches in a hypothesis space, that is restricted by rule schemata, which the user defines in terms of grammars. Keywords...

. Initially introduced in the late 1960s and early 1970s, dynamic programming algorithms have become increasingly popular in automatic speech recognition. There are two reasons why this has occurred: First, the dynamic programming strategy can be combined with avery e#cient and practical pruning strategy so that very large search spaces can be handled. Second, the dynamic programming strategy has turned out to be extremely #exible in adapting to new requirements. Examples of such requirements are the lexical tree organization of the pronunciation lexicon and the generation of a word graph instead of the single best sentence. In this paper, we attempt to systematically review the use of dynamic programming search strategies for small#vocabulary and large#vocabulary continuous speech recognition. The following methods are described in detail: search using a linear lexicon, search using a lexical tree, language-model look-ahead and word graph generation. 1 Introduction Search strategie...

Hidden Markov models were introduced in the beginning of the 1970's as a tool in speech recognition. During the last decade they have been found useful in addressing problems in computational biology such as characterising sequence families, gene nding, structure prediction and phylogenetic analysis. In this paper we propose several measures between hidden Markov models. We give an ecient algorithm that computes the measures for left-right models, e.g. prole hidden Markov models, and briey discuss how to extend the algorithm to other types of models. We present an experiment using the measures to compare hidden Markov models for three classes of signal peptides. Introduction A hidden Markov model describes a probability distribution over a potentially innite set of sequences. It is convenient to think of a hidden Markov model as generating a sequence according to some probability distribution by following a rst order Markov chain of states, called the path, from a sp...

Abstract not found

This work presents experiments to recognize pattern sequences using hidden Markov models (HMMs). The pattern sequences in the experiments are computed from speech signals and the recognition task is to decode the corresponding phoneme sequences. The training of the HMMs of the phonemes using the collected speech samples is a difficult task because of the natural variation in the speech. Two neural computing paradigms, the Self-Organizing Map (SOM) and the Learning Vector Quantization (LVQ) are used in the experiments to improve the recognition performance of the models. A HMM consists of sequential states which are trained to model the feature changes in the signal produced during the modeled process. The output densities applied in this work are mixtures of Gaussian density functions. SOMs are applied to initialize and train the mixtures to give a smooth and faithful presentation of the feature vector space defined by the corresponding training samples. The SOM maps similar feature vect...

The EuTRANS project aims at using Example-Based approaches for the automatic development of Machine Translation systems --accepting text and speech input-- for limited domain applications. During the first phase of the project, a speech translation system that is based on the use of automatically learnt Subsequential Transducers has been built. This paper contains a detailed and to a long extent self-contained overview of the transducer learning algorithms and system architecture, along with a new approach for using categories representing words or short phrases in both input and output languages. Experimental results using this approach are reported for a task involving the recognition and translation of sentences in the hotel reception communication domain, with a vocabulary of 683 words in Spanish. A translation word error rate of 1.97% is achieved in real time factor 2.7 in a Personal Computer.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.1. Approaches to Overcoming Environmental Variability . . . . . . . . . . . . . . 6 1.1.1. Re-Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.2. Multi-Style Training . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.1.3. Environmental Compensation Using Dynamic Adaptation . . . . . . . . . . 8 1.2. Towards Environment-Independent Recognition . . . . . . . . . . . . . . . . 8 1.2.1. Sources of Environmental Variability . . . . . . . . . . . . . . . . . . 9 1.2.2. Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . 9 1.3. Dissertation Outline . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Chapter 2 Overview of Environmental Robustness in Speech Recognition . . . . . . 12 2.1. Sources of Degradation...

Many conventional speech recognition systems are based on the use of hidden Markov models (HMM) within the context of discriminant-based pattern classification. While the speech recognition objective is a low rate of misclassification, HMM design has been traditionally approached via maximum likelihood (ML) modeling which is, in general, mismatched with the minimum error objective and hence suboptimal. Direct minimization of the error rate is difficult because of the complex nature of the cost surface, and has only been addressed recently by discriminative design methods such as generalized probabilistic descent (GPD). While existing discriminative methods offer significant benefits, they commonly rely on local optimization via gradient descent whose performance suffers from the prevalence of shallow local minima. As an alternative, we propose the deterministic annealing (DA) design method that directly minimizes the error rate while avoiding many poor local minima of the cost. DA is derived from fundamental principles of statistical physics and information theory. In DA, the HMM classifier's decision is randomized and its expected error rate is minimized subject to a constraint on the level of randomness which is measured by the Shannon entropy. The entropy constraint is gradually relaxed, leading in the limit of zero entropy to the design of regular nonrandom HMM classifiers. An efficient forward--backward algorithm is proposed for the DA method. Experiments on synthetic data and on a simplified recognizer for isolated English letters demonstrate that the DA design method can improve recognition error rates over both ML and GPD methods.

In this article, we describe an efficient beam search algorithm for statistical machine translation based on dynamic programming (DP). The search algorithm uses the translation model presented in Brown et al. (1993). Starting from a DP-based solution to the traveling-salesman problem, we present a novel technique to restrict the possible word reorderings between source and target language in order to achieve an efficient search algorithm. Word reordering restrictions especially useful for the translation direction German to English are presented. The restrictions are generalized, and a set of four parameters to control the word reordering is introduced, which then can easily be adopted to new translation directions. The beam search procedure has been successfully tested on the Verbmobil task (German to English, 8,000-word vocabulary) and on the Canadian Hansards task (French to English, 100,000-word vocabulary). For the medium-sized Verbmobil task, a sentence can be translated in a few seconds, only a small number of search errors occur, and there is no performance degradation as measured by the word error criterion used in this article. 1.