It was mid-June, and Finland had just entered a warm and lovely summer. The workshop organized in Nokia Arena Tampere aimed to introduce the currently popular implementation of audio and speech technology as well as foster networking with local companies. To enhance the flexibility, the workshop adopted the hybrid approach that allowed all the participants to join the event without the restriction of time and space. In the morning session, the workshop focused on two main topics, namely the audio and speech data for AI applications and machine learning methods. The content of the presentation included music and audio content analysis, embedded information content, special audio, and data collection issues. From all the presentations, we have acquired a broad understanding of the key applications and research direction of speech technology. Additionally, we also understand that it is more crucial to have a wider data diversity rather than a huge quantity in the aspect of machine learning.
After the relaxing lunch break, the workshop continued by having the case studies provided by the companies WordDive, Meluta Oy and Yousician. The presentations focused on speech and music processing. Some industrial implementations were also introduced, for instance, the music recognition in the interactive mobile application introduced by Yousician. The projects provided a more realistic point of view of audio analysis in our daily life. Finally, the workshop ended with the Associate professor Roel Pieters giving a speech about Human-Robot Communication and Post-doctoral researcher Toni Heittola introducing the MARVEL project and pilots of smart cities. The AI Hub Audio and Speech Technology workshop has been a success! The abundant amount of speech during the whole-day event enabled everyone to discuss different industry examples and expand the network. We are looking forward to meeting you in the near future!
Written by Chia-Hsin Wu, researcher of AI Hub Tampere