Mahsa Elyasi
|
Ph.D. Center for Spoken Language Understanding (CSLU) Oregon Health & Science University (OHSU) (Formerly OGI) Download my CV here. Email: ms.elyasi.l@gmail.com Find me on Google Scholar, and Linkedin. |
Research Interests
- Speech and Language Processing
- Prosody modeling/adaptation/analysis
- Voice Conversion and Text-To-Speech (TTS) Synthesis
- Machine Learning and Pattern Recognition
- Deep Neural Networks
Education
- Ph.D. in Computer Science and Engineering, Aug 2020,
Oregon Health & Science
University, Portland, OR.
Thesis: Worked on Prosody modeling and it's applications to TTS and speaker group classification under supervision of Dr. Jan P.H. van Santen. - M.S. in Computer Engineering, 2012,
Sharif
University of Technology, Tehran, IRAN.
Thesis: "Worked on the effect of phase information in speech enhancement" under supervision of Dr. Hossein Sameti. - B.S. in Computer Engineering, 2010,
University College of Nabi Akram, Tabriz, IRAN.
Positions
- Speech Research contractor at BioSpeech Inc, March 2020 - Present
- Graduate Research Assistant at CSLU, Oregon Health and Science University, Fall 2012-Present
- Research Scientist at ObEN Inc, 2016-2017
- Summer Intern at Sensory Inc, Oct 2015-Jan 2016
- Graduate Research Assistant at SPL, Sharif University of Technology, Fall 2010-2012
Publications
- M.S. Elyasi.L, J. van Santen, Recurrent Convolutional Neural Network for Classification of Speaker Groups based on Prosodic Information, WiML, 2017.
- M.S. Elyasi.L, J. van Santen, Automatic, model-based detection of pauseless phrase boundaries from fundamental frequency and duration features, 9th ISCA Speech Synthesis Workshop, 2016. [pdf]
- M.S. Elyasi.L, J. van Santen, FOOT-BASED INTONATION FOR TEXT-TO-SPEECH SYNTHESIS USING NEURAL NETWORKS, Speech Prosody (SP), 2016. [pdf]
- M.S. Elyasi.L, J. van Santen, SPEAKER INTONATION ADAPTATION FOR TRANSFORMING TEXT-TO-SPEECH SYNTHESIS SPEAKER IDENTITY, ASRU, 2015. [pdf]
- M.S. Elyasi.L, J. van Santen, S.H. Mohammadi, A. Kain, DATA-DRIVEN FOOT-BASED INTONATION GENERATOR FOR TEXT-TO-SPEECH SYNTHESIS, Interspeech, 2015. [pdf]
- M.S. Elyasi.L, J. van Santen, MODELING FUNDAMENTAL FREQUENCY DYNAMICS IN HYPOKINETIC DYSARTHRIA, SLT, 2015 [PDF]
- M.S. Elyasi.L, E. Klabbers, J. van Santen, A NOVEL PITCH DECOMPOSITION METHOD FOR THE GENERALIZED LINEAR ALIGNMENT MODEL, ICASSP, 2014. [pdf]
- M.S. Elyasi.L, H. Veisi, H. Sameti, THE EFFECT OF PHASE INFORMATION IN SPEECH ENHANCEMENT AND SPEECH RECOGNITION, ISSPA, 2012. [pdf]
- S.H. Mohammadi, H. Sameti, M.S. Elyasi.L, A. Tavanaei, KNNDIST: A Nonparametric distance measure for speaker segmentation, Interspeech, 2012. [pdf]
Reviewer
- Speech Prosody
- WiML
- Interspeech
Skills
- Languages: Python, C++, R
- ToolKits: TensorFlow, Praat, TextGrid, XML, Festival, HTS.