In this paper, the application of semi-supervised manifold learning techniques to the task of verifying hypothesized occurrences of spoken terms is investigated. These techniques are applied in a two stage spoken term detection framework where ASR lattices are first generated using a large vocabulary ASR system and hypothesized occurrences of spoken query terms in the lattices are verified in a...

This paper describes the feasibility of using semi-scripted spoken lectures as stimulus materials in a test of academic listening. The context for this study was the development of a revised test of academic listening designed to place enrolled university students into one of two levels of a language support course for non-native speakers. Because academic listening often involves listening to ...

This paper describes the data collected in Wizard of Oz experimentsin a spoken dialogue system, WAXHOLM, that provides informationon boat traffic in the Stockholm archipelago. The data consist ofutterance-length speech files, their corresponding transcriptions,and log files of the dialogue sessions. Apart from the spontaneousdialogue speech, the speech material also comp...

Recently ITC-irst (Interactive Sensory System division) and University of Pavia (Medical Informatics Labs) are working together to realize an intelligent (adaptive) dialogue system with language understanding capabilities. In this framework, some telemedicine services able to handle multimodal interactions (i.e. input/output can be provided by both voice and/or other devices such as: keyboard, ...

This paper describes how speakers adapt their language during error resolution when interacting with the animated agent Pixie. A corpus of spontaneous human-computer interaction was collected at the Telecommunication museum in Stockholm, Sweden. Adult and children speakers were compared with respect to user behavior and strategies during error resolution. In this study, 16 adults and 16 childre...

Conversations abound with uncertainties of various kinds. Treating conversation as inference and decision making under uncertainty, we propose a task independent, multimodal architecture for supporting robust continuous spoken dialog called Quartet. We introduce four interdependent levels of analysis, and describe representations, inference procedures, and decision strategies for managing uncer...

There are a number of restrictions in human-machine interactions, which continue to warrant a better control of response utterances in spoken dialogue systems. Indeed, the human user often has to deal with unnatural responses, and therefore requires some experience with such systems in order to improve interactions. This problem is re-examined here, with the aim of evaluating how human users ar...