Veranstaltung

Exploring Temporal Response Functions (TRFs) in Real-World Speech Processing

09. Dezember 2024 - 10. Dezember 2024

Join us for a workshop dedicated to exploring how the brain processes speech in real-life scenarios, leveraging the insights from Temporal Response Functions (TRFs) analysis. This event aims to delve into the complex dynamics of natural speech as encountered in everyday situations, such as casual conversations during a walk or amidst the ambient sounds of a café. Our focus is on unraveling how the brain navigates speech that is inherently dynamic—characterized by turn-taking, interruptions, and often incomplete sentences—using electroencephalograpy (EEG).

 

This workshop will address the challenges associated with capturing and analyzing EEG data and audio recordings in naturalistic settings, alongside pragmatic markers of conversation (e.g. head turning).

 

• We will discuss the critical aspects of identifying, extracting, and representing linguistic features most relevant to natural speech processing. Discussions will pivot around the best practices and best tools for extracting these linguistic elements from complex natural conversations.

• We will discuss the optimal strategies for dealing with multiple feature vectors simultaneously, a challenge inherent in processing the multidimensional data derived from real-world speech scenarios.

• We will discuss modelling non-auditory features of a conversation, such as head turning, pauses and gaze-dynamics. These considerations are vital not only for the purposes of data cleaning but also for their role as non-verbal elements within conversations, offering a more holistic view of human communication dynamics.

• We will discuss the critical importance of capturing and analyzing language properties across different temporal scales, from the immediate impact of word onset to the integrative understanding of entire paragraphs. This conversation will focus on methodologies and strategies for effectively representing these temporal dynamics in speech processing, highlighting the challenges and opportunities in modelling both the rapid and extended cognitive processes involved in natural language comprehension.

 

By addressing these topics, we hope to provide a comprehensive framework that enriches our collective understanding and capabilities in dealing with the complexities of real-life speech processing. Our goal is to foster a collaborative environment for discussing innovative methodologies, sharing (non-publish) insights, and addressing the complex questions surrounding the study of speech processing under dynamic, real-world conditions.

 

We hope that whether you are seeking to refine your research approach, explore new analytical techniques, or contribute to the broader discourse on TRFs and EEG analysis in natural speech, this workshop will offer a stimulating and enriching experience for all participants.

 

Social Event (Sunday):

Before diving into the intellectual rigor of the workshop, we will immerse ourselves in the festive joy and warmth of not just one, but two magical Christmas markets in the heart of Bremen. Begin your adventure at the historical market set around the UNESCO World Heritage-listed Town Hall, where traditions come alive amidst a picturesque setting. Savor the rich flavors of mulled wine, roasted almonds, and other seasonal delights as you wander through the stalls. Just a short walk away, experience the medieval charm of the Schlachte-Zauber on the Weser River, a promenade turned into a medieval wonderland with artisan crafts, hearty fare, and captivating riverside entertainments.

Zeitraum
09. Dezember 2024 - 09. Dezember 2024 | 09:30 Uhr - 20:00 Uhr
10. Dezember 2024 - 10. Dezember 2024 | 10:00 Uhr - 18:30 Uhr
Sprache
Englisch
Veranstaltungsort
Hanse-Wissenschaftskolleg
Lehmkuhlenbusch 4
27753 Delmenhorst
Lecture Hall
Barrierefreiheit
barrierefrei
Organisatoren
Dr. Martin Bleichner, Carl von Ossietzky Universität Oldenburg
Prof. Elana Zion Golumbic, Bar-Ilan University, Israel
Art der Veranstaltung
Workshop
Downloads
Program | pdf, 422 KB