Content
During the winter term 2011/2012 the Lehrstuhl für Informatik
6 offers the laboratory course Automatic Speech Recognition for the main studies.
In this course we will cover the following topics:
- acoustic modeling: mixtures of Gaussians and the EM-Algorithm
- language modeling: smoothing of probability distributions using absolute discounting
- search: finding a (near-)optimal hypothesis in a very large
probabilistic search space based on dynamic programming
- (a bit of) signal processing: feature extraction from audio data
These are the building blocks of any modern speech recognition system.
During the lab course, we will see how these components can be implemented
efficiently. We will develop a small digit recognition system and then apply
these techniques to larger recognition tasks.
Schedule
The first meeting will be announced by email.
You will need to register for this course here.
Assignment
The course is assigned to applied computer science or as specialization.
Enquiries
Enquiries concerning the organization of the lab course may be
sent to Martin Sundermeyer, Lehrstuhl für Informatik 6.
|