Embedded Files
Odette Scharenborg
Odette Scharenborg
Inclusive speech technology: Developing automatic speech recognition for everyone
Automatic speech recognition (ASR) is increasingly used, e.g., in emergency response centers, domestic voice assistants, and search engines. Because of the paramount relevance spoken language plays in our lives, it is critical that ASR systems are able to deal with the variability in the way people speak (e.g., due to speaker differences, demographics, different speaking styles, and differently abled users). ASR systems promise to deliver objective interpretation of human speech. Practice and recent evidence however suggests that the state-of-the-art ASRs struggle with the large variation in speech due to e.g., gender, age, speech impairment, race, and accents. The overarching goal in our research is to uncover bias in and improve fairness of ASR systems to work towards inclusive speech technology. In this talk, I will present systematic experiments aimed at quantifying, identifying the origin of, and mitigating the bias of state-of-the-art ASR systems on speech from different, typically low-resource, groups of speakers thus developing more inclusive ASR systems.
Speaker's bio
Odette Scharenborg is a Full Professor of Inclusive Speech Communication at the Delft Inclusive Speech Communication (DISC) group at the Multimedia Computing Group at Delft University of Technology, the Netherlands. Her research aims to build inclusive speech technology, i.e. making speech technology available for everyone irrespective of how they speak or what language they speak. In her research, she considers technical aspects as well as ethical and societal aspects. Her particular focus is on human and automatic speech processing, child speech, non-nativeness, low-resource languages, atypical speech, and preferably combinations of these.
From 2017-2025, Odette was on the Board of the International Speech Communication Association (ISCA), the largest international society on speech science and technology. She served as Vice-president from 2021-2023 and as President from 2023-2025. From 2018-2022, she was a member of the IEEE Speech and Language Processing Technical Committee. In 2025, she was the General Chair of Interspeech 2025, the flagship conference of ISCA and the largest international conference on speech science and technology.
Katta Spiel
Katta Spiel
Dynamic Concepts — Static Machines
Researchers and developers in the area of language technologies have recognised issues associated with minority language use, also regarding minority genders. Within the gut instinct of any good computer scientist lies then, of course, the desire to approach the issues productively and provide solutions. These have included algorithmic adjustments, suggestions for pre- and/or post-editing, data expansion (to specifically include marginalised language use in data collection and curation practices) or the synthesis of missing data. All these suggestions are driven by the imaginary of a per-se gender-inclusive utopia. This talk will interrogate this imaginary and present alternative considerations acknowledging the fundamental mismatch between the rigid constraints of any machinistic approach and the dynamic concepts (like gender) governing our social interactions. I encourage the field to reflect on the collective goals we presume and what else we could offer in world that becomes more and more polarised — with no small contribution by language technologies themselves.
Speaker's bio
Katta Spiel is an Assistant Professor for ‘Critical Access in Embodied Computing’ at TU Wien. They research marginalised perspectives on embodied computing through a lens of Critical Access. Their work informs design and engineering supporting the development of technologies that account for the diverse realities they operate in. In their interdisciplinary collaborations with disabled, neurodivergent and/or nonbinary peers, they conduct explorations of novel potentials for designs, methodologies and innovative technological artefacts.
They received their PhD in 2018 from TU Wien and after a year at KU Leuven, they conducted postdoctoral work as an FWF-Hertha Firnberg Scholar, also at TU Wien. Their work has received several international and national awards, including the SIGCHI 2020 Outstanding Dissertation Award as well as the Förderungspreis der Stadt Wien in der Sparte Mathematik, Informatik, Naturwissenschaft, Technik in 2022 and they were awarded an ERC Starting Grant in 2023. They further serve on the Austrian Monitoring Board on the UN CRPD.
Sofie Decock
Sofie Decock
Beyond the binary: Experimental insights into gender-neutral pronouns and inclusive interpreting
Recent calls for gender‑inclusive language have focused on enabling reference beyond the gender binary, most visibly through the introduction of gender‑neutral pronouns. This talk examines this ongoing shift toward inclusive nonbinary reference through a dual lens: language users’ perceptions of emerging gender-neutral pronouns and the practical challenges of implementing gender‑inclusive strategies in interpreter‑mediated communication. In a first part, I present findings from a series of survey experiments on gender‑neutral pronouns in Dutch, French and Norwegian – three languages in which such pronouns have been introduced and (partly) codified relatively recently, in contrast to more established forms such as English singular they or Swedish hen. These languages offer a unique opportunity to study language change in progress and to examine how differences in linguistic structure shape the uptake of inclusive pronominal forms. Across languages, we investigate two central uses of gender‑neutral pronouns: specific reference to nonbinary individuals and generic reference to people in general. The studies address speakers’ (evolving) familiarity with, attitudes toward and reported use of these pronouns, as well as their effects on text comprehensibility, text appreciation and mental gender representations. The second part of the talk turns to interpreter‑mediated interaction, reporting on an intervention study showing that targeted training in gender‑inclusive nonbinary language significantly improves interpreting fluency and accuracy. Together, this work provides an evidence-based foundation for discussing future directions of gender‑inclusive linguistic practices in professional communication and beyond.
Speaker's bio
Sofie Decock is Associate Professor of Applied Linguistics at the Department of Translation, Interpreting & Communication at Ghent University, Belgium. One of her main research interests centers on language and gender beyond the binary, with a particular focus on how gender‑inclusive language is used and perceived in professional communication, ranging from newspaper articles to workplace documents and interpreter‑mediated interactions. In this work, she takes into account differences in language systems and linguacultural contexts, as well as the perspectives of both L1 and L2 speakers.
She is co-editor of the journal Tijdschrift voor Taalbeheersing (since 2025), serves as a member of the Standing Committee of the Belgian American Educational Foundation (since 2022), and is also a member of the Erasmus+-funded CLADES network on Critical Language Awareness, Democratic Engagement and Sustainability (since 2024). She previously served as a board member of the Vereniging Interuniversitair Overleg Taalbeheersing (2018–2024). She has obtained FWO–SNF WEAVE funding for a research project on gender‑neutral pronouns in Dutch, French and Norwegian.
Page updated
Google Sites
Report abuse