#11.8 Interspeech 2024 Impressions - Siyang Wang

In this episode, Georgia Maniati, a Speech Scientist specializing in Text-to-Speech, shares her career path. She recounts her evolution from a linguist to a speech scientist, detailing her experiences in Greece, Edinburgh, and Italy, and how these shaped her current role. Georgia also discusses the current hurdles in text-to-speech technology, particularly for low-resource languages such as Greek.00:00:11 - intro00:00:49- Georgia's vita00:10:20 - Fellowships for young researchers00:17:04- Georgia's role at Samsung00:21:31 - hard skills in her position00:26:47 - recent, exciting projects she worked on00:31:22 - challenges in developing synthetic voices00:33:41 - Greek is a low-resource language, which challenges does it imply?00:34:38 - naturalness of TTS00:39:03 - mentoring and volunteering activities00:43:37 - young female researchers in speech00:45:20 - science communication for non expert audiences  0:53:32 - gender bias in TTS00:53:37 - current blind spots in industry00:56:08 - biggest challenges in her journey00:59:38 - biggest motivation at work and what's next01:01:11 - what did she wish to know earlier in her career?01:02:33 - the question from Titouan Parcollet01:03:49 - her question for the next guest01:04:19 - outro

The most long-awaited episode of 2025, the mental health episode is here. In this episode the Speech Pitch team asked your question to Jasmina Bakic, who is a scientific researcher and psychologist in a psychology practice person-centered in Amsterdam. You will hear 3 stories from our audience and Jasmina's responses to them and to the many questions related to burn out, self doubt, toxic enviroments, self-harm, ADHD and how to seek for help.Hosts: Priyanshi Pal, Spyretta Leivaditi, Orchid Chetia Phukan, Sarthak JainEditors: Pascal Hecker, Snigdha Banik, Spyretta Leivaditi, Kalliopi Kakamouka0:00:00 intro0:00:x introducing Jasmina and her areas of interest at work0:04:52 first story: burn out0:15:35 potential way outs for burnout0:17:50 second story: depression upon entering a PhD0:23:15 should you share mental health concerns with your supervisor?0:25:25 if your advisor is not empathetic, how should you get help instead?0:27:55 third story: how do I build effective relationships with my co-workers0:37:50 how do you ensure better collaboration in the research effort?0:43:42 short question section: how can one stay focused if your supervisor doesn't support you0:47:30 how can one cope with toxic behaviour of people around you?0:57:00 how do you know that resigning is the right decision?1:00:49 how to get out of a phase of self-doubt?1:04:56 what is a healthy way of self-worth?1:11:28 Balance between competition and your own pace1:14:57 fear of being judged to ask "stupid questions"1:19:35 how can you react to unfruitful feedback?1:22:02 how to overcome the feeling of not being sufficient1:25:56 how to spot someone who might be considering self-harm1:29:40 how to persuade someone to seek help1:32:08 how can a supervisor grief after an unfortunate event1:36:40 ADHD in adults1:43:08 take-home-message1:43:48 outro

This is the uncut version of our episode with Say IT Labs. We meet Lukas Latacz and Erich Reiter, founders of Say IT Labs, and discuss their career, the founding process of their company, and their contribution to pathological speech using their Artificial Intelligence (AI) applications.Hosts: Pascal Hecker, Spyretta LeivaditiEditors: Pascal Hecker, Snigdha Banik, Spyretta Leivaditi, Kay Berkling, Kalliopi Kakamouka, Mohammed Mosuily0:00:00 - Intro0:00:30 - Introducing Erich Reiter and Lukas Latacz0:04:22 - The origin story of SAY IT Labs0:05:56 - How Erich came to Belgium0:07:33 - Challenges when founding the company0:12:18 - How did they secure initial funding0:15:00 - SAY IT Labs as a company0:19:43 - The science of stuttering and the game stutter stars0:26:10- How is it validated?0:30:00 - Different languages0:31:35 - Smart glasses for Parkinson's0:38:30 - Advice for young entrepreneurs0:47:20 - Work life balance0:53:08 - Question from Titouan: What is your biggest challenge to overcome for your product to succeed?0:56:44 - Question for the next guest: How to transition from monolingual to multilingual approaches?0:57:06 - Support SAY IT Labs: interns0:57:50: - Outro

In this episode we meet Lukas Latacz and Erich Reiter, founders of SAY IT Labs, and discuss their career, the founding process of their company, and their contribution to pathological speech using their Artificial Intelligence (AI) applications.Hosts: Pascal Hecker, Spyretta LeivaditiEditors: Pascal Hecker, Snigdha Banik, Spyretta Leivaditi, Kalliopi Kakamouka, Orchid Chetia Phukan, Sarthak Jain, Vishakha Choudhary

Titouan Parcollet is a “Research Scientist at the Samsung AI Center Cambridge” and an “adjunct researcher at the Cambridge Machine Learning Systems Lab from the University of Cambridge”. Further, he is an “Associate Professor on leave from the Laboratoire Informatique d'Avignon (LIA) and Avignon Université (FR)”. His current Research focus is on self-supervised / representation learning and on continual learning. He played an instrumental part in the development of SpeechBrain and Pytorch-Kaldi. In this episode you will follow Titouan’s origin story, how he entered university, his PhD journey, his teaching approach and of course his current research topics. Hosts: Spyretta Leivaditi, Pascal Hecker Editors: Pascal Hecker, Janice Huang and Snigdha Banik 00:00:00 - Intro 00:00:20 - Welcoming Titouan Parcollet 00:01:01 - Titouan's entry into university and early career 00:05:48 - Titouan's PhD journey 00:15:33 - PhD exchange with the Mila institute 00:17:58 - Importance of a PhD advisor and how to chose your PhD position 00:27:50 - Titouan's teaching approach and experience 00:35:50 - His current research topic and his view on the field 00:40:42 - His view on academia and industry 00:43:26 - Work-life balance, mental health, burnout 00:54:34  -Imposter syndrome 01:03:07 - The SpeechBrain toolkit 01:16:06 - The Flower framework 01:17:55 - The E-SSL project: Efficient Self-Supervised Learning for Inclusive and Innovative Speech Technologies 01:22:51 - Titouan answer's Florian Eyben's question 01:24:41 - Titouan's question for the next guest 01:25:49 - Outro

Florian Eyben spearheads technology and innovation at audEERING, focusing on developing industry-leading products for speech emotion recognition and deep learning-based audio analysis. He earned my PhD in Computational Paralinguistics from TUM in Munich, Germany. He also specialize in deep learning, audio feature extraction, signal processing, project management, and tech innovation. He is the lead author of the openSMILE toolkit and a co-author of the GPU-accelerated LSTM-RNN training toolkit, CuRRENNT. In this podcast episode, Florian shares his academic and professional journey, gives insights about openSMILE and of course shares how audEERING was founded. Enjoy !! Hosts: Pascal Hecker, Spyretta Leivaditi Editors: Janice Huang and Pascal Hecker Chapters: 00:00:00 Intro00:00:26 Welcoming Florian Eyben00:00:55 Florian's background and research journey00:10:46 More about openSMILE00:22:37 Founding audEERING and what it is00:35:10 For young researchers00:50:30 Encouragement from Florian01:00:50 Fun questions01:17:37 Outro

Meet Rob van Son, Senior researcher at the Netherlands Cancer Institute Amsterdam who shares his interests and his impressions on Interspeech 2024 in Kos.Host: Zhengjun Yue

Meet Shrikanth Narayanan who is professor of Electrical and Computer Engineering and Niki & C. L. Max Nikias Chair in Engineering shares his research interests and his impressions on Interspeech 2024 in Kos. Host: Orchid Chetia Phukan

Meet Shekhar Nayak, associate professor in Speech Technology in Campus Fryslân of University of Groningen. Listen about his interests, his academic journey and of course his impression on Interspeech 2024 in Kos. Host: Spyretta Leivaditi

Siyang is a PhD student in KTH Royal Institute of Technology shares his experience and impression of Interspeech 2024 in Kos. Host: Paige Tuttosi