I am looking to hire hourly
student research assistants for the fall
quarter for three
different projects.
1) Automatic speech recognition for low-resource languages: Current speech recognition technology relies on machine learning methods that leverage large amounts of transcribed data. In this project, we are interested in languages other than English, in which case there may not be a lot of data available. This project will involve harvesting data from the web for augmenting small data sets and learning methods that account for style differences.
2) Automatic detection and classification of marine mammals: Detection of marine mammals is important for avoiding sonar operations when they are present and for tracking populations. The problem is challenging because of the highly variable noise conditions in the oceans. We have developed some prototype algorithms in Matlab and are looking for someone to work on porting these algorithms to C/C++ and factoring the implementation to leverage parallel processing and allow experimentation with large amounts of data. Once the software is in place, the goal is to explore semi-supervised learning algorithms.
3) Speaking style analysis: Spoken language carries much more information than one can see in the transcript, including emotion, strength of belief, sarcasm, etc. In this project, we investigate signal processing methods to extract cues that an automatic system could use to recognize this information.
All projects will require computer implementation of algorithms for data processing and analysis.
1) Automatic speech recognition for low-resource languages: Current speech recognition technology relies on machine learning methods that leverage large amounts of transcribed data. In this project, we are interested in languages other than English, in which case there may not be a lot of data available. This project will involve harvesting data from the web for augmenting small data sets and learning methods that account for style differences.
2) Automatic detection and classification of marine mammals: Detection of marine mammals is important for avoiding sonar operations when they are present and for tracking populations. The problem is challenging because of the highly variable noise conditions in the oceans. We have developed some prototype algorithms in Matlab and are looking for someone to work on porting these algorithms to C/C++ and factoring the implementation to leverage parallel processing and allow experimentation with large amounts of data. Once the software is in place, the goal is to explore semi-supervised learning algorithms.
3) Speaking style analysis: Spoken language carries much more information than one can see in the transcript, including emotion, strength of belief, sarcasm, etc. In this project, we investigate signal processing methods to extract cues that an automatic system could use to recognize this information.
All projects will require computer implementation of algorithms for data processing and analysis.
Qualification/expertise:
Computer programming skills and
experience, including: experience with unix,
programming languages such as C, C++ or Java and
scripting languages such as Perl or Python.
Students should have completed and done well in
both CS142 and 143 (at least). Knowledge of
probability is very useful for anticipated work
on machine learning. For projects 2 & 3,
experience with Matlab is desirable.
Hours & Salary:
The position will
involve 10-19 hrs/week, flexible hours, depending
on student availability. The salary is $12-15/hr,
depending on experience.
Application Documents:
CV,
(unofficial) transcript, and the contact
information of one or two references,
including either the instructor for one of
your programming courses or a supervisor for
whom you've done programming.
Contact:
Interested candidates may
send their application material to Prof.
Ostendorf at ostendor@u.washington.edu.
If you are interested in the first position,
please copy Aaron Jaech (ajaech@uw.edu),
if interested in the second position, please
copy Nicole Nichols (nmn3@u.washington.edu),
and if interested in the third position, please
copy Vicky Zayats (vzayats@uw.edu).
