The thoughtful elephant: strategies for spoken dialog systems

Abstract

In this paper we present technology used in spoken dialog systems for applications of a wide range. They include tasks from the travel domain and automatic switchboards as well as large scale directory assistance. The overall goal in developing spoken dialog systems is to allow for a natural and flexible dialog flow similar to human-human interaction. This imposes the challenging task to recognize and interpret user input, where he/she is allowed to choose from an unrestricted vocabulary and an infinite set of possible formulations. We therefore put emphasis on strategies that make the system more robust while still maintaining a high level of naturalness and flexibility. In view of this paradigm, we found that two fundamental principles characterize many of the proposed methods: 1) to consider available sources of information as early as possible, and 2) to keep alternative hypotheses and delay the decision for a single option as long as possible.We describe how our system architecture caters to incorporating application specific knowledge, including, for example, database constraints, in the determination of the best sentence hypothesis for a user turn. On the next higher level, we use the dialog history to assess the plausibility of a sentence hypothesis by applying consistency checks with information items from previous user turns. In particular, we demonstrate how combination decisions over several turns can be exploited to boost the recognition performance of the system. The dialog manager can also use information on the dialog flow to dynamically modify and tune the system for the specific dialog situations. An important means to increase the "intelligence" of a spoken dialog system is to use confidence measures. We propose methods to obtain confidence measures for semantic items, whole sentences and even full N-best lists and give examples for the benefits obtained from their application. Experiences from field tests with our systems are summarized that have been found crucial for the system acceptance.Index Terms-Application specific knowledge, combined decisions, confidence measures, dialog history, natural language understanding, spoken dialog systems.

On Combining Language Models: Oracle Approach

“…Although it has turned out to be a rather difficult task to beat the (almost) standard class/word n-grams (typically Ò ¾ or ¿), there has been a great deal of interest in grammar based language models [1]. A promising approach for limited domain applications is the use of semantically motivated phrase level stochastic context free grammars (SCFGs) to parse a sentence into a sequence of semantic tags which are further modeled using Ò-grams [2,9,10,3].…”

Statistical language model adaptation: review and perspectives

“…The idea behind this approach is that well-understood parts of a sentence occur in most hypotheses of an N -best list, whereas for misrecognitions different candidates usually appear in different hypotheses. Thus, the effect of a recognition error is distributed over several competing hypotheses and does not result in a strong error reinforcement (Souvignier et al, 2000). A variation on the same concept was recently presented in (Gretter and Riccardi, 2001), based on exploiting sausages rather than N -best lists.…”

Voice access of global information for broad-band wireless: technologies of today and challenges of tomorrow

“…For example, sub-dialogues can be used for the confirmation of the understood concepts, error recovery, reduction or expansion of the scope of the user's request, clarification of the ambiguities, etc. [105]- [107]. The response to the user is finally formulated as sentences and produced as speech signals to be transmitted to the user.…”