Prosody Modifications for Question-Answering in Voice-Only Settings
Preprint
Abstract
Many popular form factors of digital assistant---such as Amazon Echo, Apple Homepod or Google Home---enable the user to hold a conversation with the assistant based only on the speech modality. The lack of a screen from which the user can read text or watch supporting images or video presents unique challenges. In order to satisfy the information need of a user, we believe that the presentation of the answer needs to be optimized for such voice-only interactions. In this paper we propose a task of evaluating usefulness of prosody modifications for the purpose of voice-only question answering. We describe a crowd-sourcing setup where we evaluate the quality of these modifications along multiple dimensions corresponding to the informativeness, naturalness, and ability of the user to identify the key part of the answer. In addition, we propose a set of simple prosodic modifications that highlight important parts of the answer using various acoustic cues.Keywords
All Related Versions
This publication has 7 references indexed in Scilit:
- Perception of Sentence Stress in Speech Correlates With the Temporal Unpredictability of Prosodic FeaturesCognitive Science, 2015
- Predictability affects early perceptual processing of word onsets in continuous speechNeuropsychologia, 2011
- Why Um Helps Auditory Word Recognition: The Temporal Delay HypothesisPLOS ONE, 2011
- Signal-based and expectation-based factors in the perception of prosodic prominenceLaboratory Phonology, 2010
- Prosody-driven Sentence Processing: An Event-related Brain Potential StudyJournal of Cognitive Neuroscience, 2005
- On the Role of Sentence Stress in Sentence ProcessingLanguage and Speech, 1977
- Estimating the Reliability, Systematic Error and Random Error of Interval DataEducational and Psychological Measurement, 1970