Focus of Attention in the Hearsay-II Speech Understanding System.

摘要

Hearsay-2 is a complex, distributed-logic system for speech understanding developed at Carnegie-Mellon University. Processing is performed by independent, data-directed knowledge source processes that examine and alter values in a global data base representing hypothesized phonetic segments, syllables, words, and phrases, as well as the hypothetical temporal and logical relationships among them. The question of how the numerous potential activities of the knowledge sources should be scheduled to complete the interpretation of an utterance in minimal time is called the 'focus of attention problem.' Near optimal focusing is especially important in a speech understanding system because of the very large solution space that potentially needs to be searched. Thus, this focus of attention problem is representative of general resource allocation problems involving cooperative and competitive processes. Using the concepts of stimulus and response frames of scheduled knowledge source instantiations, competition among alternative responses, goals, and the desirability of a knowledge source instantiation, a general attentional control mechanism is developed. This general focusing mechanism facilitates the experimental evaluation of a variety of specific attentional control policies (such as best-first, bottom-up, and top-down search strategies) and allows the modular addition of specialized heuristics for the speech understanding task. Empirical results demonstrate the effectiveness of the focusing principles, and possible directions for future research are considered. (Author)