Tanel Alumäe
E-mail: tanel.alumae@phon.ioc.ee
Address:
Foneetika ja kõnetehnoloogia labor
TTÜ Küberneetika Instituut
Akadeemia tee 21
12618 Tallinn
Estonia
Telephone: +372 620 4201
Here is my CV.
Link to my CV at ETIS (Estonian Research Portal).
Research Interests
- Speech recognition, particularly language modelling
- Spoken document retrieval
- Dialogue systems
- Natural language processing
- Machine learning, statistical methods
Current Activities
- Speech Recognition (part of the Estonian National Programme for Language Technology, 2011-2017)
Past Activities
- Research and development of methods for Estonian speech recognition (part of the Estonian National Programme for Language Technology, 2006-2010)
Theses supervision
We are looking for Master and Phd students! There are many available topics.
Some Publications
- Slaidid riigi IT-juhtide seminaril
- Slides of my talk Advanced Rich Transcription System for Estonian Speech at Baltic HLT 2018.
- Slides of my presentation at the conference “Audiovisual translation – new trends in translation technology”
- Modeling under-resourced languages for speech recognition. Language Resources and Evaluation, 2016. Joint work with Mikko Kurimo, Seppo Enarvi, Ottokar Tilk, Matti Varjokallio, André Mansikkaniemi.
- LSTM for Punctuation Restoration in Speech Transcripts. Interspeech 2015, Dresden, Germany. Joint work with Ottokar Tilk. {Code}
- Multi-Domain Recurrent Neural Network Language Model for Medical Speech Recognition. Baltic HLT 2014, Kaunas, Lithuania. Joint work with Ottokar Tilk.
- Full-duplex Speech-to-text System for Estonian. Baltic HLT 2014, Kaunas, Lithuania.
- Neural network phone duration model for speech recognition. Interspeech 2014, Singapore. {Code}
- Recent improvements in Estonian LVCSR. SLTU 2014, Saint Petersburg, Russia.
- Phone duration modeling using clustering of rich contexts. Interspeech 2013, Lyon, France. Joint work with Rena Nemoto. [Poster]
- Multi-domain neural network language model. Interspeech 2013, Lyon, France. [Poster]
- Transcription system for semi-spontaneous Estonian speech. Baltic HLT 2012, Tartu, Estonia.
- Maximum entropy language model adaptation for mobile speech input. Interspeech 2012, Portland, OR, USA. Joint work with Kaarel Kaljurand. [Poster]
- Open and extendable speech recognition application architecture for mobile environments. SLTU 2012, Cape Town, South Africa. Joint work with Kaarel Kaljurand.
- TSAB - web interface for transcribed speech collections. Interspeech 2011, Firenze, Italy. Joint work with Ahti Kitsik.
- Efficient estimation of maximum entropy language models with N-gram features: an SRILM extension. Interspeech 2010, Chiba, Japan. Joint work with Mikko Kurimo. [Poster]
- Domain adaptation of maximum entropy language models. ACL 2010, Uppsala, Sweden. Joint work with Mikko Kurimo. [Poster]
- Comparison of different modeling units for language model adaptation for inflected languages. Proceedings of CICLing 2008, Haifa, Israel, pp. 488-499. [Poster]
- LSA-based language model adaptation for highly inflected languages. Interspeech 2007. Antwerp, Belgium, pp. 2357-2360.
- Automatic compound word reconstruction for speech recognition of compounding languages. Proceedings of NODALIDA 2007, pp. 5-12.
- Methods for Estonian Large Vocabulary Speech Recognition. Ph. D. Thesis, Tallinn University of Technology, 2006.
- Sentence-adapted factored language model for transcribing Estonian speech. Proceedings of ICASSP 2006. Toulouse, France, pp. 429–432.
- Large vocabulary continuous speech recognition for Estonian using morpheme classes. Proceedings of Interspeech 2004 - ICSLP, pp. 389–392.
See my CV for a full list of publications.