This paper describes connectionist techniques for recognition of Broadcast News. The fundamental difference between connectionist systems and more conventional mixture-of-Gaussian systems is that connectionist models directly estimate posterior probabilities as opposed to likelihoods. Access to posterior probabilities has enabled us to develop a number of novel approaches to confidence estimation, pronunciation modelling and search. In addition we have investigated a new feature extraction technique based on the modulation-filtered spectrogram (MSG), and methods for combining multiple information sources. We have incorporated all of these techniques into a system for the transcription

A digital VLSI implementation of a psychoacoustically and psysiologically motivated preprocessing algorithm is introduced. The algorithm consists of several stages motivated by the signal processing in the human ear and was successfully applied to a wide range of applications by the Medical Physics Group at the University of Oldenburg. The processing scheme has been partitioned and will be implemented in two ASICs. It is shown how an application of the algorithm has been used to determine the necessary wordlengths for a transfer of the floating point version of the model into a version suitable for a hardware implementation.