Vector Microprocessors by Krste Asanovic Doctor of Philosophy in Computer Science University of California, Berkeley Professor John Wawrzynek, Chair Most previous research into vector architectures has concentrated on supercomputing applications and small enhancements to existing vector supercomputer implementations. This thesis expands the body of vector research by examining designs appropriate for single-chip full-custom vector microprocessor implementations targeting a much broader range of applications. I present the design, implementation, and evaluation of T0 (Torrent-0): the first single-chip vector microprocessor. T0 is a compact but highly parallel processor that can sustain over 24 operations per cycle while issuing only a single 32-bit instruction per cycle. T0 demonstrates that vector architectures are well suited to full-custom VLSI implementation and that they perform well on many multimedia and human-machine interface tasks. The remainder of the thesis contains ...

This report is concerned with integrating connectionist networks into a hidden Markov model (HMM) speech recognition system, This is achieved through a statistical understanding of connectionist networks as probability estimators, first elucidated by Herve Bourlard. We review the basis of HMM speech recognition, and point out the possible benefits of incorporating connectionist networks. We discuss some issues necessary to the construction of a connectionist HMM recognition system, and describe the performance of such a system, including evaluations on the DARPA database, in collaboration with Mike Cohen and Horacio Franco of SRI International. In conclusion, we show that a connectionist component improves a state of the art HMM system. ii Part I INTRODUCTION Over the past few years, connectionist models have been widely proposed as a potentially powerful approach to speech recognition (e.g. Makino et al. (1983), Huang et al. (1988) and Waibel et al. (1989)). However, whilst connec...

Over the past 40 years, significant progress has been made in the fields of speech recognition and speech understanding. Current state-of-the-art speech recognition systems are capable of achieving word-level accuracies of 90 % to 95 % on continuous speech recognition tasks using 5000 words. Even larger systems, capable of recognizing 20,000 words are just now being developed. Speech understanding systems have recently been developed that perform fairly well within a restricted domain. While the size and performance of modern speech recognition and understanding systems are impressive, it is evident to anyone who has used these systems that the technology is primitive compared to our own human ability to understand speech. Some of the difficulties hampering progress in the fields of speech recognition and understanding stem from the many sources of variation that occur during human communication. One of the sources of variation that occurs in human communication is the different ways that words can be pronounced. There are many causes of pronunciation variation, such as: the phonetic environment in which the word occurs, the dialect of the speaker,

This paper describes the architecture and performance of the Berkeley Restaurant Project (BeRP), a medium-vocabulary, speaker-independent, spontaneous continuous speech understanding system currently under development at ICSI. BeRP serves as a testbed for a number of our speech-related research projects, including robust feature extraction, connectionist phonetic likelihood estimation, automatic induction of multiplepronunciation lexicons, foreign accent detection and modeling, advanced language models, and lip-reading. In addition, it has proved quite usable in its function as a database frontend, even though many of our subjects are non-native speakers of English. 1 OVERVIEW The BeRP system functions as a knowledge consultant whose domain is restaurants in the city of Berkeley, California. As a knowledge consultant, it draws inspiration from earlier consultants like VOYAGER [15]. Users ask spoken language questions of BeRP, which directs questions to the user and then queries a dat...

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

Parallelism and distribution have been considered the key features of neural processing. The term parallel distributed processing is even used as a synonym for artificial neural networks. Nevertheless, the actual implementations are still in search of the appropriate model to "naturally represent" neural computing. And the final judgement is always given in performance figures -- keeping the parallelization issue high on the neurosimulation agenda. Two approaches have yielded the best results: parallel simulations on general-purpose computers, and specially developed neurohardware. Programming neural networks on parallel machines requires high-level techniques reflecting both inherent features of neuromodels and characteristics of the underlying computers. On the other hand, emulation of the neuroparadigm requires that the functioning of neural operations be mimicked directly by the hardware. Both approaches are presented, and their advantages and shortcomings are outlined.

Previously, we have demonstrated that feed-forward networks may be used to estimate local output probabilities in hidden Markov model (HMM) speech recognition systems. Here these connectionist techniques are integrated into the DECIPHER system, with experiments being performed using the speaker independent DARPA RM database. Our results indicate that: . connectionist probability estimation can improve performance of a context independent maximum likelihood trained HMM system, . performance of the connectionist system is close to what can be achieved using (context dependent) HMM systems of much higher complexity, and . mixing connectionist and maximum likelihood estimates can improve the performance of a state-of-theart context dependent HMM system. 1 INTRODUCTION Previous investigations, both theoretical and experimental, have indicated that feed-forward networks (typically, multilayer perceptrons, MLPs) may be used to estimate local HMM output probabilities [1, 6]. Our previous p...

This paper describes the research underway for the ESPRIT WERNICKE project. The project brings together anumber of different groups from Europe and the US and focuses on extending the state-of-the-art for hybrid hidden Markov model/connectionist approaches to large vocabulary, continuous speech recognition. This paper describes the specific goals of the research and presents the work performed to date. Results are reported for the resource managementtalker-independent recognition task. The paper concludes with a discussion of the projected future work.

We report on recent advances in the ICSI connectionist speech recognition project. Highlights include: . Experimental results showing that connectionist methods can improve the performance of a context independent maximum likelihood trained HMM system, resulting in a performance close to that achieved using state of the art context dependent HMM systems of much higher complexity. . Mixing (context independent) connectionist probability estimates with maximum likelihood trained context dependent models to improve the performance of a state of the art system . The development of a network decomposition method that allows connectionist modelling of context dependent phones efficiently and parsimoniously, with no statistical independence assumptions. . L&H Speechproducts, Ieper, B-8900 Belgium. y. SRI International, Menlo Park CA 94025, USA. Part I INTRODUCTION The dominant approach to automatic continuous speech recognition is statistical [5, 7]. The resulting methods, which use cru...

This paper describes the architecture and performance of the Berkeley Restaurant Project (BeRP), a medium-vocabulary, speaker-independent, spontaneous continuous speech understanding system currently under development at ICSI. BeRP serves as a testbed for a number of our speech-related research projects, including robust feature extraction, connectionist phonetic likelihood estimation, automatic induction of multiplepronunciation lexicons, foreign accent detection and modeling, advanced language models, and lip-reading. In addition, it has proved quite usable in its function as a database frontend, even though many of our subjects are non-native speakers of English. 1 OVERVIEW The BeRP system functions as a knowledge consultant whose domain is restaurants in the city of Berkeley, California. As a knowledge consultant, it draws inspiration from earlier consultants like VOYAGER [15]. Users ask spoken language questions of BeRP, which directs questions to the user and then queries a dat...