Partner event: ARSID Spring Topical Seminar

”Some challenges in learning to speak Danish” by John Tøndering, University of Copenhagen.

Foreigners who learn Danish often reach high proficiency, enabling full participation in Danish society. However, some pronunciation features may remain challenging even after decades of residence. Rather than deficits, these features reflect unique attributes of Danish - its flat intonation, vowel reductions, and soft articulation. In this talk, I will explore how some of these pronunciation patterns shape communication in Danish.

“Cracking the Code: Unveiling the Influence of Opaque Sound Structures on Language Learning” by Fabio Trecca, Aarhus University

Children acquire their native language(s) by tracking regularities in, and extracting patterns from, the linguistic input they receive from caregivers and peers. The surface properties of a language, including its constituent sounds, syllable structure, and stress patterns, form a large pool of cues from which children can deduce and generalize linguistic rules. However, spoken languages vary significantly in transparency, affecting the availability of these surface properties for computation. Spoken Danish, for instance, is characterized by pervasive phonetic reductions, resulting in a relatively opaque pronunciation compared to closely related languages. In this talk, I will discuss how being exposed to a language with an opaque sound structure can hinder language processing and
learning, providing evidence and discussing potential implications for education and intervention strategies.

“The visual modality of language: gesture and speech in interaction” by Maria Graziano, Lund University Humanities Lab.

Most research on language mainly focuses on speech. However, face-to-face communication is multimodal. Indeed, speakers often gesture when they talk, no matter what language they speak and what culture they belong to. In this talk, I will give an overview on the relationship between speech and gesture and the function of gestures in discourse.

“Audio-visual speech enhancement” by Daniel Michelsanti, Aalborg University

Speech is perhaps the most common way that people use to communicate with each other. Often, this kind of communication is harmed by several sources of disturbance that may have different nature, such as the presence of competing speakers, the loud music during a party, and the noise inside a car cabin. In this talk, I will explore how the use of artificial intelligence algorithms can reduce the impact of such sources of disturbance in speech communication by exploiting visual information of the speaker (i.e. lip reading).

“Speech visualisation for Computer-Aided Pronunciation Training” by Jeppe R. Frisvad, DTU

Compute Visual display of spectrograms as well as 2D and 3D computer animations of the lips and oral cavity can be beneficial for pronunciation learning. However, computer-aided pronunciation training (CAPT) systems often suffer from difficulties in giving learners adequate, accurate feedback and an inability to provide accurate diagnosis of pronunciation errors. We believe more advanced speech visualisation
techniques have the potential to significantly improve the state of current CAPT systems. We believe that speech visualisation is useful for generating interactive visual feedback that can help new language learners improve their pronunciation. This talk explores options and potentials of speech visualisation in the field of CAPT.