Links & FAQ

I would like announce the 2011 Microsoft Research PhD Fellowship Program. Each fellowship includes a $28,000 stipend, payment of tuition/fees, and conference travel for two years. In addition, we offer each Fellow the opportunity to be a 12-week, salaried intern at Microsoft Research. The Microsoft Research PhD Fellowship Program for 2011 has started accepting applications beginning today and the deadline for submission of applications is Friday, October 8, 2010, at 5:00 P.M. Pacific Time. Applicants must be enrolled in a full-time PhD program in an accredited university in the US or Canada with a research focus in Computer Science, Mathematics, or Electrical Engineering. For additional details of eligibility requirements and the application procedure, please see:http://research.microsoft.com/en-us/collaboration/awards/apply-us.aspx

Microsoft Research is also pleased to begin accepting applications for the 2011 Microsoft Research Graduate Women's Scholarship Program, today, August 17, 2010. Each scholarship includes a $15,000 stipend, payment of tuition/fees, and conference travel for two years. The deadline for submission of applications is Friday, October 8, 2010, at 5:00 P.M. Pacific Time. Applicants must be full-time, female students, enrolled in their first year of graduate program in an accredited university in the US or Canada with a research focus in Computer Science, Mathematics, or Electrical Engineering. For qualifications and application procedures, please seehttp://research.microsoft.com/en-us/collaboration/awards/fellows-women.aspx.

Q15: What is difference between training, test, and development data set?

A15: Data split:

1) Training data set (or called as enrollment data set). Say you may have 3 classes, each class may have 2 samples. For speaker verification, we will say each enrollment speaker has two sessions.

2) Test data set

3) Development data set (or background data set): the goal of this set is to find optimal parameter which can be applied to "Test data set"

A good data split should make sure:

there is no overlap among the above 3 datasets.

Development data set should close match what we encounter in "Test data set", if we can.

Q16: What is the impact factor of major speech conferences

A16:

As for conference impact factors, please find an indication as given by http://citescholar.org as follows (2010 Impact Factor, Conference):

1.16 ICASSP,

1.05 INTERSPEECH

Q17: Where can I find public corpus for speech Enhancement/identification?

A17: Here is one example from MITLL (some researchers also used it for speaker identification)

https://www.ll.mit.edu/mission/cybersec/HLT/corpora/SpeechCorpora.html

Lincoln Laboratory Speech Enhancement Corpus (LLSEC) - The LLSEC was collected under DARPA funding to further research in the area of speech enhancement. The LLSEC is a mutli-microphone speech corpus collected under different scenarios to realistically reflect a wide variety of degraded speech conditions.
- View report: LLSEC corpus
- Download: Version 1.0 of LLSEC (166 MB)

Q18: Where to find the file lists for NIST SRE?

A18:

http://cls.ru.nl/~saeidi/file_library/I4U.tgz

http://code.google.com/p/prism-set/

Q19: Is there any toolkit for telephone speech enhancement?

A19:

Speech processing, transmission and quality aspects (stq), distributed speech recognition, advanced front-end feature extraction algorithm, compression algorithm,” in ETSI standard document-ETSI ES 202 050 v1.1.1, 2002.

http://www.etsi.org/deliver/etsi_es/202000_202099/202050/01.01.05_60/es_202050v010105p.pdf

http://www.etsi.org/deliver/etsi_es/202000_202099/202050/01.01.05_60/es_202050v010105p0.zip

Q20: Upcoming Signal Processing Conferences

A20:

http://www.signalprocessingsociety.org/conferences/upcoming-conferences/

Q21: How to write a great research paper?

A21: Please refer: How to write a great research paper (by Simon Peyton Jones)

1.Don’t wait: write

2.Identify your key idea

3.Tell a story

4.Nail your contributions

5.Related work: later

6.Put your readers first (examples)

7.Listen to your readers

More: http://research.microsoft.com/~simonpj

Q22: Where to find free accent corpus:

A22:

http://accent.gmu.edu/about.php