Other translations of this page: None.

Opening for a PhD position

Update: the position is filled.

The Laboratory of Phonetics and Speech Technology seeks qualified candidates for one PhD student position in the field of automatic speech recognition. The preemptive title of the thesis is Adaptive Deep Neural Networks for Speech Recognition and Keyword Search.

We offer a fulltime post for a period of two to a maximum of four years. Salary will depend on education and experience.

Topic description

Speech recognition models are known to be fragile against mismatches in training and testing conditions. The mismatches can be caused by the environment and channel or the physical and social characteristics of the speaker. Although DNNs improve speech recognition performance in both clean and noisy conditions, processing data that is to some extent distinct from training data is still a major research problem. This thesis investigates adapting DNN models for speech recognition to conditions different from the training data, using little or no additional human-transcribed speech data. We will seek methods that can be applied in an online fashion, without requiring long-running steps to retrain the existing models. The focus is on developing DNN models that are adaptive to the input signal, i.e., that adapt to the varying and changing speech conditions automatically. The methods developed during the thesis will be applied for speech recognition and keyword search. The experiments concentrate on speech data that has been previously been difficult to process, such as interviews made with a far standing microphone, recordings from historical multimedia archives.

Institution

Laboratory of Phonetics and Speech Technology
Institute of Cybernetics
Tallinn University of Technology

Requirements

Candidates should have obtained a graduate degree in computer science. Experience in speech processing is appreciated.

The successful candidate should fulfill the following criteria:

  • Degree in computer science, electrical engineering or a discipline with a related background.
  • Excellent programming skills (C/C++, Python)
  • Experience with Linux and bash-scripting
  • Very good math background
  • Very good oral and written English communication skills
  • Interest in speech recognition research

Applying

Each application should include:

  • CV including a list of relevant research experience and a list of publications (if applies)
  • Transcript of records BSc/MSc
  • Letter of motivation
  • Names of two references.
  • Any other supporting information or documents.

Applications should be sent no later than June 15 2016 to: tanel.alumae@phon.ioc.ee. Shortlisted candidate will undergo a series of tests including technical reading and writing in English and programming.

The position is fully funded depending on the qualification and professional experience of the successful candidate. The position is full time for two years with a possible extension. The starting date is September 1st 2016.

 
studying/dnn-phd.en.txt · Last modified: 2016/08/30 16:22 by tanel