معرفی
Helene ZGANIC is a Researcher affiliated with Multispeech, a research team common to Inria, Loria (Universite de Lorraine/CNRS/Inria), and Vivoka. Her work focuses on speech-to-text (STT) systems, natural language understanding (NLU), voice data protection, and acoustic diagnostics, often involving collaborations between academic institutions and industry partners.
Key projects she has contributed to include Acoust.IA, designed for sound diagnostic tools, and the VoicePrivacy Challenge, which explores voice anonymization techniques. She has also been involved in the COMPRISE project, addressing weakly supervised learning for speech and text processing, and the Oriental Language Recognition (OLR) challenge, highlighting her expertise in automatic language recognition systems.
Her research emphasizes reducing reliance on manually labeled datasets, improving voice assistant technologies, and advancing ethical AI applications in audio engineering and collaborative systems. Academic-industrial partnerships, such as with Vivoka, underscore her commitment to translating research into practical solutions.
