Rafid Antoon Sukkar

Rejection of non-digit strings for connected digit speech recognition

A high reliability digit string recognizer/rejection system that processes spoken words through an HMM recognizer to determine a string of candidate digits, a filler model for each digit in the digit string, and other information. Next, a weighted sum is generated for each digit in the string and for a filler model for each digit in the string. A confidence score is generated for each digit by subtracting the filler weighted sum from the digit weighted sum. The confidence score for each digit is then compared to a threshold and, if the confidence score for any of the digits is less than the threshold, the entire digit string is rejected. If the confidence scores for all of the digits in the digit string are equal to or greater than the threshold, then the candidate digit string is accepted as a digit string.

Verifying and correcting recognition string hypotheses using discriminative utterance verification

Abstract Utterance verification (UV) is a process by which the output of a speech recognizer is verified to determine if the input speech actually includes the recognized keyword(s). The output of the speech verifier is a binary decision to accept or reject the recognized utterance based on a UV confidence score. In this paper, we extend the notion of utterance verification by presenting an utterance verification method that will be utilized to perform three tasks: (1) detect non-keyword strings (false alarms), (2) detect keyword substitution errors, and (3) selectively correct substitution errors when N -best string hypotheses are available. The utterance verification method presented here employs a set of verification-specific models that are independent of the models used in the recognition process. The verification models are trained using a discriminative training procedure that seeks to minimize the verification error by simultaneously maximizing the rejection of non-keywords and misrecognized keywords while minimizing the rejection of correctly recognized keywords. The error correction is performed by reordering the hypotheses produced by an N -best recognizer based on a UV confidence score.

A Framework for Event Correlation in Communication Systems

This paper introduces a framework for event correlation in communication systems. We will show how the concept of a class in objectoriented methodology can be used to provide scalability to the framework. Events and system topology information are combined to generate the causal information needed for correlation. Geometric representation of codewords is used to overcome the noise factor. Temporal reasoning is explored to reduce noise and increase the number of event patterns that can be detected. The framework has been applied to a wireless communication system.

Reducing computational complexity and response latency through the detection of contentless frames

In this paper, we present a method that manipulates the decoding network to reduce both computational complexity and response latency while maintaining high ASR accuracy. The method employs a TSVQ (tree structured vector quantization) classifier that reliably discriminates between silence and non-silence frames. Reductions in computational complexity and response latency are achieved through three techniques: 1) silence skipping, 2) silence-based pruning of the dynamic programming network, and 3) early decision. Experimental results on a connected digit task and a large vocabulary company name task show that the proposed method can reduce ASR response latency by more than 82%. Furthermore, the computational complexity, measured in CPU seconds, was reduced by 13.6% on the connected digit task and 6.7% on the company name task while maintaining the recognition accuracy of the baseline system.

Unified speech recognition for the landline and wireless environments

Compared to the landline network environment, the wireless environment presents new factors affecting ASR performance (or accuracy). Our goal here is to determine these factors, and their relative importance, and then to devise methods to mitigate them. We approach this goal, first, by conducting a set of experiments where we use a state of the art ASR system trained on landline speech data, and then compare its performance in landline network conditions to its performance across a variety of wireless network conditions. Based on the results of these experiments, we determine critical factors affecting ASR accuracy. We then use multi-condition acoustic models to mitigate these factors and show that the resulting ASR system is able to not only achieve high accuracy across the various wireless network conditions, but also maintain its high accuracy across landline network conditions. This leads to a recognition system that is channel independent which is a very desirable property for telecom based applications.

Joint MCE estimation of VQ and HMM parameters for Gaussian mixture selection

Vector quantization (VQ) has been explored in the past as a means of reducing likelihood computation in speech recognizers which use hidden Markov models (HMMs) containing Gaussian output densities. Although this approach has proved successful, there is an extent beyond which further reduction in likelihood computation substantially degrades the recognition accuracy. Since the components of the VQ frontend are typically designed after model training is complete, this degradation can be attributed to the fact that VQ and HMM parameters are not jointly estimated. In order to restore the accuracy of a recognizer using VQ to aggressively reduce computation, joint estimation is necessary. We propose a technique which couples VQ frontend design with minimum classification error training. We demonstrate on a large vocabulary subword task that in certain cases, our joint training algorithm can reduce the string error rate by 79% compared to that of VQ mixture selection alone.