Publications

A new training procedure called multi-style training has been developed to improve performance when a recognizer is used under stress or in high noise but cannot be trained in these conditions. Instead of speaking normally during training, talkers use different, easily produced, talking styles. This technique was tested using a speech data base that included stress speech produced during a workload task and when intense noise was presented through earphones. A continuous-distribution talker-dependent Hidden Markov Model (HMM) recognizer was trained both normally (5 normally spoken tones) and with multi-style training (one token each from normal, fast, clear, loud, and question-pitch talking styles). The average error rate under stress and normal conditions fell by more than a factor of two with multi-style training and the average error rate under conditions sampled during training fell by a factor of four.

Artificial neural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech and image recognition. These models are composed of many nonlinear computational elements operating in parallel and arranged in patterns reminiscent of biological neural nets. Computational elements or nodes are connected via weights that are typically adapted during use to improve performance. There has been a recent resurgence in the field of artificial neural nets caused by new net topologies and algorithms, analog VLSI implementation techniques, and the belief that massive parallelism is essential for high performance speech and image recognition. This paper provides an introduction to the field of artificial neural nets by reviewing six important neural net models that can be used for pattern classification. These nets are highly parallel building blocks that illustrate neural net components and design principles and can be used to construct more complex systems. In addition to describing these nets, a major emphasis is placed on exploring how some existing classification and clustering algorithms can be performed using simple neuron-like components. Single-layer nets can implement algorithms required by Gaussian maximum-likelihood classifiers and optimum minimum-error classifiers for binary patterns corrupted by noise. More generally, the decision regions required by any classification algorithm can be generated in a straightforward manner by three-layer feed-forward nets.

Substantial improvements in speech recognition performance on speech produced under stress and in noise have been achieved through the development of techniques for enhancing the robustness of a base-line isolated-word Hidden Markov Model recognizer. The baseline HMM is a continuous-observation system using mel-frequency cepstra as the observation parameters. Enhancement techniques which were developed and tested include: placing a lower limit on the estimated variances of the observations; addition of temporal difference parameters; improved duration modelling; use of fixed diagonal covariance distance functions, with variances adjusted according to perceptual considerations; cepstral domain stress compensation; and multi-style training, where the system is trained on speech spoken with a variety of talking styles. With perceptually-motivated covariance and a combination of normal (single-frame) and differential cepstral observations, average error rates over five simulated-stress conditions were reduced from 20% (baseline) to 2.5% on a simulated-stress data base (105-word vocabulary, eight talkers, five conditions). With variance limiting, normal plus differential observations, and multi-style training, an error rate of 1.8% was achieved. Additional tests were conducted on a data base including nine talkers, eight talking styles, with speech produced under two levels of motor-workload stress. Substantial reductions in error rate were demonstrated for the noise and workload conditions, when multiple talking styles, rather than only normal speech, were used in training. In experiments conducted in simulated fighter cockpit noise, it was shown that error rates could be reduced significantly by training under multiple noise exposure conditions.