Publications

In this paper we examine the integration of speaker identification/verification technology into two dialogue systems developed at MIT: the Mercury air travel reservation system and the Orion task delegation system. These systems both utilize information collected from registered users that is useful in personalizing the system to specific users and that must be securely protected from imposters. Two speaker recognition systems, the MIT Lincoln Laboratory text independent GMM based system and the MIT Laboratory for Computer Science text-constrained speaker-adaptive ASR-based system, are evaluated and compared within the context of these conversational systems.

This article describes our progress on automated, two-way English-Korean translation of text and speech for enhanced military coalition communications. Our goal is to improve multilingual communications by producing accurate translations across a number of languages. Therefore, we have chosen an interlingua-based approach to machine translation that readily extends to multiple languages. In this approach, a natural-language-understanding system transforms the input into an intermediate-meaning representation called a semantic frame, which serves as the basis for generating output in multiple languages. To produce useful, accurate, and effective translation systems in the short term, we have focused on limited military-task domains, and have configured our system as a translator's aid so that the human translator can confirm or edit the machine translation. We have obtained promising results in translation of telegraphic military messages in a naval domain, and have successfully extended the system to additional military domains. The system has been demonstrated in a coalition exercise and at Combined Forces Command in the Republic of Korea. From these demonstrations we learned that the system must be robust enough to handle new inputs, which is why we have developed a multistage robust translation strategy, including a part-of-speech tagging technique to handle new works, and a fragmentation strategy for handling complex sentences. Our current work emphasizes ongoing development of these robust translation techniques and extending the translation system to application domains of interest to users in the military coalition environment in the Republic of Korea.

Experiments demonstrated that sigmoid multilayer perceptron (MLP) networks provide slightly better risk prediction than conventional logistic regression when used to predict the risk of death, stroke, and renal failure on 1257 patients who underwent coronary artery bypass operations at the Lahey Clinic. MLP networks with no hidden layer and networks with one hidden layer were trained using stochastic gradient descent with early stopping. MLP networks and logistic regression used the same input features and were evaluated using bootstrap sampling with 50 replications. ROC areas for predicting mortality using preoperative input features were 70.5% for logistic regression and 76.0% for MLP networks. Regularization provided by early stopping was an important component of improved performance. A simplified approach to generating confidence intervals for MLP risk predictions using an auxiliary "confidence MLP" was developed. The confidence MLP is trained to reproduce confidence intervals that were generated during training using the outputs of 50 MLP networks trained with different bootstrap samples.

Pattern-classification and clustering algorithms are key components of modern information processing systems used to perform tasks such as speech and image recognition, printed-character recognition, medical diagnosis, fault detection, process control, and financial decision making. To simplify the task of applying these types of algorithms in new application areas, we have developed LNKnet-a software package that provides access to more than 20 pattern-classification, clustering, and feature-selection algorithms. Included are the most important algorithms from the fields of neural networks, statistics, machine learning, and artificial intelligence. The algorithms can be trained and tested on separate data or tested with automatic cross-validation. LNKnet runs under the UNM operating system and access to the different algorithms is provided through a graphical point-and-click user interface. Graphical outputs include two-dimensional (2-D) scatter and decision-region plots and 1-D plots of data histograms, classifier outputs, and error rates during training. Parameters of trained classifiers are stored in files from which the parameters can be translated into source-code subroutines (written in the C programming language) that can then be embedded in a user application program. Lincoln Laboratory and other research laboratories have used LNKnet successfully for many diverse applications.