Modeling pronunciation variants is an important topic for automatic speech recognition. This paper investigates the pronunciation modeling at the lexical level, and presents a detailed modeling of the probabilities of the pronunciation variants. The approach is evaluated on the French ESTER2 corpus, and a significant word error rate reduction is achieved through the use of context and speaking ...

This paper considers finite-blocklength achievability for rate-compatible codes. For a fixed number of messages, random coding analysis determines a sequence of achievable error probabilities for a sequence of blocklengths. However, traditional random coding achievability draws each code independently so that it does not show that a family of rate-compatible codes can achieve that sequence of e...

In this paper we search for the optimum tests that minimize the sum of two error probabilities.

This chapter is dedicated to scope of the application of Importance Sampling Techniques to the design phase of Neyman-Pearson Neural Detectors. This phase usually requires the application of MonteCarlo trials in order to estimate some performance parameters. The classical Monte-Carlo method is suitable to estimate high event probabilities but not suitable to estimate very low event probabilitie...

In a Bayesian network, a convergence error may arise in the probabilities computed for a node with two or more parents in the same loop (in this paper such a node is called a convergence node), when in the computation of these probabilities the information of the parents is processed as if they are independent. In this paper the range of the convergence error for the basic case of a loop with j...

We explore the connection between channel and source coding problems. Our study centers on the class of binning schemes that characterize the largest known achievable rates for channel and source coding problems with partial side information and multiterminal scenarios. The binning scheme can be viewed as a set of bins with each bin forming a channel code (with error probability 22) and set of ...

We present a lower bound on the probability of symbol error for maximum-likelihood decoding of lattices and lattice codes on a Gaussian channel. The bound is tight for error probabilities and signal-to-noise ratios of practical interest, as opposed to most existing bounds that become tight asymptotically for high signal-to-noise ratios. The bound is also universal: it provides a limit on the hi...

The realization, two decades ago, that quantum mechanics can be a powerful resource to speed up important computational tasks [1] led to intense research efforts to find adequate physical systems for quantum computation. One of the hurdles to a viable technology is the requirement to prepare, manipulate, and measure quantum bits (qubits) with near perfect accuracy: Imperfect control leads to er...