University Homepage School Homepage School Contacts School Search

Automatic Speech Recognition (ASR) -2009/10-

[course descriptor]

Lecturers

News

• There will be a tutorial on shell programming on Wednesday, 10th March between 14:00-14:50 at AT 4.12.
• The first lecture: Monday, 11th January 2010, 15:00-15:50, at AT 4.12
  
• In addition to "Speech Processing" as the pre-requisite course, it would be advantageous that you take either "Introductory Applied Machine Learning" (IAML), or "Probabilistic Modelling and Reasoning" (PMR) in advance. "Machine Learning and Pattern Recognition" (MLPR) and "Machine Translation" (MT) are related second semester courses, along with the PPLS course on "Speech Synthesis".
• You can have a look at the last year's couse web page

Schedule

• The lectures will be held on Mondays and Thursdays at 15:00-15:50 (in Semester 2) at AT 4.12.

Schedule

Week	Monday (AT 4.12)		Thursday (AT 4.12)
1	11.Jan	Signal analysis (HS)	14.Jan	Signal analysis (HS)
2	18.Jan	Statistical PR (HS)	21.Jan	LAB (HS)
3	25.Jan	Statistical PR (HS)	28.Jan	HMMs 1 (SR)
4	01.Feb	HMMs 2 (SR)	04.Feb	LAB (HS)
5	08.Feb	Modelling speech with HMMs 1 (SR)	11.Feb	Modelling speech with HMMs 2 (SR)
6	15.Feb	Speaker adaptation 1 (SR)	18.Feb	LAB (HS)
7	22.Feb	Speaker adaptation 2 (SR)	25.Feb	LAB (HS)
8	01.Mar	Lexicon and language model (SR)	04.Mar	LAB (HS)
9	08.Mar	Search and decoding (SR)	11.Mar	Discrimination and feature combination (SR)
10	15.Mar	Environmental robustness (HS)	18.Mar	Guest lecture: the AMI large vocabulary system (Mike Lincoln)
11	22.Mar	(no lecture)	25.Mar	(no lecture)

Assignment

• See the handout for lab-session(4) and lab-session(5)
• The submission deadline is Thursday, 25th March 2010 at 16:00.
• Report templates (LaTeX) (MS Word)
• Submission instructions

Handouts

• Lecture 1 and 2 (11, 18 Jan): signal analysis handout (1) [R1.2: 18/01/10]
• Lecture 3 and 4 (18, 25 Jan): statistical pattern recognition handout (2) [R1.1: 18/01/10]
• Lecture 5 and 6 (28 Jan, 1 Feb): hidden Markov models handout (3) (slides)
• Lecture 7 and 8 (8, 11 Feb): modelling speech with HMMs handout (4) (slides)
• Lecture 9 and 10 (15, 22 Feb): speaker adaptation handout (5) (slides)
• Lecture 11 (1 Mar): lexicon and language model handout (6) (slides)
• Lecture 12 (8 Mar): search and decoding handout (7) (slides)
• Lecture 13 (11 Mar) discrimination and feature combination handout (8) (slides)
• Lecture 14 (15 Mar) robustness handout (9) (slides)
• Lecture 15 (18 Mar) The AMI ASR system slides

• lab-session (1) [R1.1: 21/01/10]
• lab-session (2) [R1.1: 04/02/10]
• lab-session (3) [R1.2: 24/02/10]
• lab-session (4) [R1.4: 03/03/10]
• lab-session (5) [R1.1: 04/03/10]
• notes on decoding parameters [R1.1: 17/03/10]
• bug reports
• Shell script examples: run the 'init-t3' command as usual in your ASR directory, you will see the new directory, 'ShellScriptExamples'.

References

• Useful texts