Interspeech 2019 is the annual conference of the International Speech Communication Association and the world's largest event on the science and technology of spoken language. A record number of visitors is expected at the 20th conference at the Messecongress Graz. More than 2000 experts from over 50 countries have signed up, including representatives from companies such as Alibaba, Alphabet (Google), Amazon, Apple, Asapp, DataBaker, Didi, Facebook, Microsoft, Naver, and Nuance Communications. The reason for the high level of interest is also the choice of topics. In addition to speech communication and signal processing, major research areas are artificial intelligence and machine learning.
Variety of applications and languages
Selected topics include solutions for the automatic translation of spoken language, new algorithms for detecting emotions and drowsiness via voice recordings, clinical voice analysis, voice recognition of public figures and the automatic recognition of fake voices or the protection of privacy in connection with the use of virtual assistants and smart speakers.
A strong scientific interest focusses on linguistic varieties. Gernot Kubin, head of the Signal Processing and Speech Communication Laboratory at TU Graz, makes the conference motto "Crossroads of Speech and Language" tangible using an example: "Regional dialects and foreign accents cause considerable problems for voice assistant systems. On the one hand, there is currently still too little data available here to fully exploit the potential of the systems. On the other hand, research has not yet found any simple approaches that would allow users to train the devices themselves."
Kubin brought Interspeech 2019 to Graz and is responsible for the conference organisation. He and his team are supported by the University of Graz, the Medical University of Graz, the University of Music and Performing Arts Graz, and the University of Maribor in Slovenia. "In terms of linguistic diversity, the European Union is the world scientific leader. Graz, in turn, has a proven interdisciplinary focus across universities in this field. From information engineering to acoustics to linguistics, many disciplines at the Graz universities are dealing with this topic," says Kubin, explaining the advantages of the venue.
In addition to professional expertise, many new event formats attract guests to Graz. Based on keynotes and survey talks by top-class speakers summarising the current state of knowledge, poster presentations, tutorials, a specialist exhibition and show & tells in which scientific prototypes and pilot systems are presented, await the participants. Those who cannot make it to Graz, can acquire knowledge at other satellite events in Budapest, Ljubljana, Stockholm and Vienna.
Event highlights for budding scientists
There are also special event offers for the approximately 600 early-stage researchers taking part in the conference. The first Interspeech Hackathon brings together international and regional students who then work together on innovative solutions for given tasks.
At the Young Female Researchers in Speech Workshop, bachelor's and master's students can network with leading women researchers from academia and industry, exchange ideas and receive motivation to take up doctoral studies. Mentoring events for doctoral students as well as an advanced training event for high-school teachers from Austria and Slovenia in cooperation with the University College of Teacher Education Styria round off the programme.
Gernot Kubin expects the conference to have "many positive effects, especially since the conference is now moving Graz onto the map of internationally renowned research locations in the fields of speech communication and artificial intelligence. In addition, the domestic industry can take advantage of the potential of the conference. On the one hand, to network more closely with the industry leaders in the digital market. And on the other hand, to attract promising young talent in the fields of AI and data science."