PROGRAM SPEECH TECH DAY 2024

Theme: towards inclusive speech technology in research and applications

As the quality of speech and voice technology has improved dramatically in recent years, there is an increased demand for using it in practical scenarios in a variety of application domains, from interaction with conversational agents and robots via tools in education and e-health to transcription services for individuals and organisations. 

The speech and voice technology community in The Netherlands is growing. Bringing this community together forms an excellent opportunity to engage with the newest insights on speech tech from academia, to hear about what companies and governmental organisations are working on, to share experiences and unite people working on speech and voice technology around shared values of inclusivity, innovation and shared purpose.

(preliminary program)

Session I Keynote

09:30 - 10:00 Walk-in
10.00 - 10:05 Welcome by Odette Scharenborg (Delft University of Technology)
10:05 - 10:50 Keynote by Dr. Olya Kudina (Delft University of Technology) | Voice-based interfaces: Diversity and inclusion considerations.

Session II  Talks from Public private sector

Session chair: Odette Scharenborg
10:50 - 11:10 Badal Marhé and Jorik van der Hoek (Achmea / Interpolis) | Voice technology with Achmea
11:10 - 11:30 Arjan van Hessen (NOTAS / University of Twente) | ASR: What else is there to wish for?

11:30 - 11:50 Short break

11:50 - 12:10 Berend Jutte (Attendi) | Advancing healthcare with language technologies: challenges and opportunities
12:10 - 12:30 Hubert van Beusekom (TNO) | About GPT-NL
12:30 - 12:45 Q&A on the basis of questions from the audience

12:45 - 13:45 Lunch

Session III Talks from Academia

Session chair: Henk van den Heuvel
13:45 - 14:05 Dragoș Bălan (University of Twente) | Dutch Open Speech Recognition Benchmark
14:05 - 14:25 Matt Coler (University of Groningen) | Fostering diversity through speech technology
14:25 - 14:45 Grzegorz Chrupała (Tilburg University) | Putting Natural in Natural Language Processing
14:45 - 15:05 Martijn Bentum (Radboud University) | What can we learn from deep models
15:05 - 15:20 Q&A on the basis of questions from the audience

15:20 - 15:40 Short break

Session IV Posters and information market

15:40 - 16:30 Visit the posters and information market

Posters:

Yun Hao, Reihaneh Amooie, Wietse de Vries, Martijn Wieling -- Utilizing Self-Supervised Learning Representations for Low-Resource Dutch Acoustic-to-Articulatory Inversion

 Alianda Lopez -- The Timing Bottleneck: Why Timing and Overlap Are Mission-Critical for Conversational User Interfaces, Speech Recognition and Dialogue Systems 

Reihaneh Amooie, Yun Hao, Jelske Dijkstra, Matt Coler, Wietse de Vries, Martijn Wieling -- Evaluating ASR Architectures for Frisian

Marcio Fuckner, Sophie Horsman, Iskaj Janssen -- Uncovering Bias in ASR Systems: Evaluating the performance of Wav2vec2 and Whisper for Dutch speakers 

Marijn Schraagen, Hans Marien -- Classifying vocabulary knowledge from speech

Tanvina Patel and Odette Scharenborg -- Improving End-to-End Models for Children’s Speech Recognition

Bastiaan Tamm, Jakob Poncelet, Mara Barberis, Maaike Vandermosten, Hugo Van hamme -- Weakly Supervised Training Improves Flemish ASR of Non-Standard Speech

Yuanyuan Zhang, Aaricia Herygers, Tanvina Patel, Zhengjun Yue, Odette Scharenborg -- Exploring Data Augmentation in Bias Mitigation Against non-native-accented Speech

16:30 - 18:00 Drinks