بورسیه دکتری برق، الکترونیک، مخابرات، کامپیوتر،فناوری اطلاعات در سوئد

بورسیه دکتری برق، الکترونیک، مخابرات، کامپیوتر،فناوری اطلاعات در سوئد

Department information

The positions are hosted by the department of Speech, Music and Hearing (TMH).

TMH is an internationally distinguished group within speech technology and associated fields dating back to the 1950s. Research at the department is truly multi-disciplinary including linguistics, phonetics, auditory perception, vision, experimental psychology, signal processing and machine learning. Rooted in an engineering modeling approach, its research forms a solid base for developing multimodal human-computer interaction systems in which speech, music, sound and gestures combine to create human-like communication.

Job description

The most successful efforts in virtually all areas of current core speech technology research involve large quantities of data combined with signal processing and machine learning. The Ph.D. positions are supported by a number of externally funded research projects in this area starting in 2016 and 2017, ranging from research around data collection, selection and processing to novel machine learning methods and applications.

Within this greater scope, TMH is hiring up to 2 Ph.D. students in two connected but distinct areas.*

* Utilization of found data in speech technology and interaction research

Within speech technology, the term “found data” is used for data that was not recorded or collected specifically in order to underlie speech technology research. There are many sizeable sources of speech and language data available to us (e.g. on the Internet and in national archives), but these are very rarely used – neither for speech technology development, nor for other research purposes – since they behave in a much less predictable manner than do data that is purpose-collected. At the same time, these data are both more representative of real-world data and more interesting for fundamental research, in the sense that they are themselves often subject of research.

This Ph.D. work involves method development for processing of and learning from large quantities of diverse found speech and language data on the one hand, and fundamental research on the data itself on the other. The latter part of the work is highly interdisciplinary and collaborative work together with researchers from a variety of disciplines.

* Development of machine learning methods for modelling language acquisition

Speech technology traditionally uses expensive linguistic resources, such as large collections of annotated speech data and lexical databases, in conjunction with machine learning methods to build models of speech recognition, understanding and synthesis. This framework lacks the flexibility of human language learning in many aspects, for example: – humans do not need annotations to learn to speak, – learning continues during the whole life, so that new words can seamlessly be added to the lexicon, – learning is inherently multimodal, so that the meaning of new words is grounded in the specific multisensory context – production and perception are strongly coupled during learning so that humans learn both from listening and speaking in the interaction with other humans.

This Ph.D. work involves modelling some of the above aspects of human language learning in order to make speech technology more flexible and adaptive to unpredictable situations. The candidate will work in a highly stimulating international environment in collaboration with researchers at the forefront in novel machine learning methodology.

Qualifications

To be admitted to the announced positions in the Ph.D. education in Speech Communication and Technology, you must have a master’s degree (or an equivalent 4-year university degree). The area is multidisciplinary and candidates from a variety of backgrounds outside of pure speech technology will be considered, including sensor technology and hardware; systems and programming; statistics and  modelling; linguistics, cognitive and behavioural science, computer vision and machine learning. The applicant should have an excellent command of spoken and written English. Applicants must be strongly motivated for doctoral studies, possess the ability to work independently and perform critical analysis and also possess good levels of cooperative and communicative abilities.

A good understanding of signal processing and machine learning is seen as a strong merit. For “Development of machine learning methods for modelling language acquisition” this is fundamental and the candidate will be expected to master both theoretical and practical aspects of machine learning methods. Furthermore, experience with any of the following is seen as an additional merit: spoken dialogue systems, speech technology, computer vision, teaching, user studies and experimental design.

Trade union representatives

You will find contact information to trade union representatives at KTH:s webpage.

Application

Application shall include the following documents:

1. Curriculum vitae.
2. Transcripts from University/ University College.
3. Document that describes and confirms the applicant’s practical experience of work in related areas.
4. Brief description of which of the areas that fits best and why.
5. Brief description of why the applicant wishes to become a doctoral student.

Please observe that all material needs to be in English, apart from the official document.

Log into KTH’s recruitment system in order to apply to this position. You are the main responsible to ensure that your application is complete according to the ad.

Your complete application must be received at KTH no later than the last day of application, midnight CET/CEST (Central European Time/Central European Summer Time).

Others

For more information about doctoral studies at KTH, visit https://www.kth.se/en/studies/phd/doctoral-studies-phd-1.9318.

We firmly decline all contact with staffing and recruitment agencies and job ad salespersons.

Disclaimer: In case of discrepancy between the Swedish original and the English translation of the job announcement, the Swedish version takes precedence.