Development and evaluation of a spoken dialogue system for academic document retrieval with a focus on reply generation

Abstract

SUMMARYA spoken dialogue system for retrieving academic documents was constructed and evaluated. In order to realize speech reply with its prosodic features properly controlled to express dialogue focus, a concept-to-speech conversion scheme was developed where the reply sentence concept was finally converted step-by-step to phone and prosodic symbol strings to be fed to a speech synthesizer, instead of simply using a text-to-speech conversion scheme as done in other conventional spoken dialogue systems. The main research focuses were on the following three points: (1) generating a reply sentence easily understood by users and unambiguous to users by controlling the amount of information included in it: deleting redundant information and adding necessary information; (2) realizing functions necessary for increasing system usability; (3) generating a speech reply with natural-sounding prosody, with the dialogue focuses correctly conveyed. Evaluations of the original system, formerly developed, gave us guidelines to realize points (1) and (2). Based on them, a revised system was constructed and evaluated, showing the validity of its new functions. It was also shown that speech replies with prosodic features representing dialogue focuses could bring about clearer information transmission.

A spoken dialogue system for document information retrieval utilizing topic knowledge

“…The retrieval module categorizes the input retrieval words in retrieval categories (items within the document information data to be matched with retrieval words), determines the type of combination of retrieval words (AND retrieval or OR retrieval), and generates the retrieval condition in a string form [8]. Document information retrieval is performed according to the generated retrieval string, and the relevant documents sorted according to the relevance score values of each document are output as the result.…”

“…These comments indicate that support depending on the user's mastery of the system should be investigated, and this will be an important future research topic. Whether to give the dialogue initiative to the user or system has already been investigated for this system [8], and a mechanism for switching the dialogue initiative has already been implemented. Therefore, we believe a relatively easy improvement concerning this point could be implemented by eliciting the user's level of experience, for example, by asking a question such as "Is this the first time you are using the system?"…”

“…Therefore, we developed a spoken dialogue system for document information retrieval tasks, which has been designed to smoothly carry out retrieval dialogues by using speech replies to present appropriate retrieval support information according to the dialogue progress conditions [8].…”

Prosodic focus control in reply speech generation for a spoken dialogue system of information retrieval

“…For each word class code in the sequence, a word entry is selected by referring to the information stack. The detail of the process is explained in [1].…”

“…From this viewpoint, we already have realized a Concept-ToSpeech (CTS) conversion scheme (a scheme to convert reply content to reply speech without a step to generate surface sentences in orthographic transcription) and succeeded in generating reply speech with its prosodic features properly controlled to express the dialogue focuses in our spoken dialogue system on document retrieval [1]. However, the rules of focusing were very simple and several problems were pointed out from users during the trial use of the system.…”