Spoken language consists of a complex, time-varying signal that contains sequential patterns that can be described in terms of statistical relations among language units. Previous research has suggested that a domain-general ability to learn structured sequential patterns may underlie language acquisition. To test this prediction, we examined the extent to which implicit sequence learning of pr...

This paper discusses issues in evaluating spoken language dialogue systems in terms of technical performance and end-user acceptance. Recent efforts in this domain have been carried out in the framework of two major research initiatives: the European Esprit long-term project Spoken Language Dialogue Systems and Components Best practice in development and evaluation (DISC) and the US American DA...

For telecommunications companies or banks, etc processing spontaneous lanaguage in helpdesk scenarios is important for automatic telephone interactions. However, the problem of understanding spontaneous spoken language is difficult. Learning techniques such as neural networks have the ability to learn in a robust manner. Recurrent networks have been used in neurocognitive or psycholinguisticall...

Connectionist theories of language propose that written language deficits arise as a result of damage to semantic and phonological systems that also support spoken language production and comprehension, a view referred to as the "primary systems" hypothesis. The objective of the current study was to evaluate the primary systems account in a mixed group of individuals with primary progressive ap...

Spoken language consists of a complex, time-varying signal that contains sequential patterns that can be described in terms of statistical relations among language units. Previous research has suggested that a domain-general ability to learn structured sequential patterns may underlie language acquisition. To test this prediction, we examined the extent to which implicit sequence learning of pr...

Abstract We introduce dGSLM, the first “textless” model able to generate audio samples of naturalistic spoken dialogues. It uses recent work on unsupervised unit discovery coupled with a dual-tower transformer architecture cross-attention trained 2000 hours two-channel raw conversational (Fisher dataset) without any text or labels. show that our is speech, laughter, and other paralinguistic sig...

We investigate the impact of disfluencies on the task of classifying natural human-human conversations into topics. Disfluencies are distinctive to spoken language, and their effect on a number of spoken language understanding tasks, including spoken language classification, remains largely unknown. We use a subset of Switchboard-I annotated for disfluencies and topics, and investigate the effe...