Leeuwarden, Netherlands

Speech Technology

when 14 May 2023 - 17 May 2023
language English
duration 1 week
credits 1 EC
fee EUR 95

This is a hands-on summer school on speech technology for anyone interested in speech tech, from absolute beginners to experts. The event is comprised mainly of interactive workshops hosted by cutting-edge start-ups, including:

Voice Synthesis: Artificial voices, speech cloning, talking avatars in video games, and more.
ReadSpeaker: The best text-to-speech for you! Quality evaluation of hyper-realistic human voices, and how ReadSpeaker creates bespoke voices to match customer requirements. Students will evaluate audio along several criterion and learn about the nuances of measures to evaluate speech quality.

Descript: Clone your own voice in just a few minutes. Learn about cutting edge voice cloning techniques, discuss ethical implications, and clone your own voice with software developed by professors at Berkeley at the company Lyrebird, since acquired by Descript. [hybrid]

8D Games: Speech technology with serious gaming. Learn about the use of speech technology in Serious Gaming for support of engagement and learning experiences in healthcare & education.

Speech Recognition: Recognizing human voices with robots, real-time transcription, and more.
LuxAI: Embodied voice-based interaction with QTrobots. Students work in groups of up to 5 people per robot and implement interactional capabilities based on tutorial templates provided by the trainer. Examples include making the robot look at the speaker when they are talking, producing “back-channeling”, recognizing emotions, and using posture-tracking and face-recognition.

Amberscript: Real-time speech recognition! - Workshop description coming soon

Explainers + Lectures
Nuance: Learn about the history of speech technology from a scholar at Nuance. From the beginnings of speech recognition and the earliest attempts at synthesis, up until the present. This presentation will be an interactive crash-course in the field suitable for beginners. [hybrid]

MSc Voice Technology: Find out about the new graduate program in voice technology at the University of Groningen and meet the program director for an interactive Q&A session.

Raspberry pi for speech tech: Get experience working on a raspberry pi in small groups on activities relating to speech tech.

Google (Daan van Esch, Senior Technical Program Manager with the Google Speech team): Building Language Technologies across the World’s Languages

Respeecher (Anna Bulakh) — an Emmy-award winning Ukrainian voice synthesis company

University of Oxford (Prof. Jieun Kiaer)

... and more: Google, Ukrainian speech tech, DIY speech activities!

Plus you will have a chance to present your own posters in a poster session attended by academics, partners from industry, and fellow students.

Note that this summer school occurs simultaneous with the annual LITHME COST Action H2020 meeting at the same place. We are organizing it strategically so participants in the summer school can attend opening/closing events, keynote lectures, lunches, and other content, thus ensuring a diversity of channels to explore speech technology suitable for different backgrounds and interests of all participants.

Course leader

Matt Coler

Target group

This school is intended for:

undergraduates interested in AI, computing science, human-machine interaction, and applied linguistics who want to learn more about the field
graduates and PhDs interested in speech tech or AI who want to learn about current issues in the field, new applications, and are interested in the professional opportunities in the field.
all students who want to network, share ideas in a poster session, and learn about the field of speech technology.

Course aim

After this summer school you will have:

1. learned about key concepts relating to the state-of-the-art in speech technology (specifically speech recognition and synthesis)
2. learned about applications of speech technology in industry (including robotics, video games, artificial voices, among others)
3. gained hands-on experience moving from theory to practice by working on diverse speech technology projects in workshops developed by academia and industry.
4. worked together in international, intercultural teams of students from different backgrounds

Credits info

1 EC
Preparation: 2 hours
Lectures: 9.3 hours per day x 3 days
Presentation: [depends on if student is presenting a poster]

Fee info

EUR 95: for UG/ENLIGHT (excluding farewell dinner)
EUR 135: for others (excluding farewell dinners)