Topic: Lectures | ELEC-E5510 - Speech Recognition D, Lecture, 25.10.2023-8.12.2023

In this list, the 2022 slides will be replaced by the 2023 ones after each lecture is given at latest. The titles may be identical, but the contents are improved each year based on feedback. The project works and their schedule changes each year.

For practicalities, e.g. regarding to the Lecture Quizzes and Exercises, check MyCourses > Course Practicalities

Select activity Lectures 1-2 Feature extraction and modeling

Lectures 1-2 Feature extraction and modeling
Select activity Lectures 1-2 slides (2023)
Lectures 1-2 slides (2023) File PDF
course organization
what is ASR
features of speech
MFCC
GMM
DNN
Select activity Lecture 1-2 exercise: Gaussian mixture model
Not available unless any of: You are a(n) Student ...
Not available unless any of:

You are a(n) Student

You are a(n) Teacher

You are a(n) Teacher (MC)
Lecture 1-2 exercise: Gaussian mixture model Assignment

Instructions can be found in the pdf file. Please upload your answer here, e.g. as a photo, text or pdf file
Select activity Lectures 3-4 - Phoneme modeling

Lectures 3-4 - Phoneme modeling
Select activity Lecture 3-4 slides (2023)
Lecture 3-4 slides (2023) File PDF
Phonemes
HMMs
Select activity Lecture 3: exercise Forward
Not available unless any of: You are a(n) Student ...
Not available unless any of:

You are a(n) Student

You are a(n) Teacher (MC)

You are a(n) Teacher
Lecture 3: exercise Forward Assignment

Please type or upload your calculations here, e.g. as a photo, text or pdf file to earn a lecture activity point.
Select activity Lecture 3: exercise Viterbi
Not available unless any of: You are a(n) Student ...
Not available unless any of:

You are a(n) Student

You are a(n) Teacher (MC)

You are a(n) Teacher
Lecture 3: exercise Viterbi Assignment

Please type or upload your calculations here, e.g. as a photo, text or pdf file to earn a lecture activity point.
Select activity Lectures 5-6 - Language Modeling

Lectures 5-6 - Language Modeling
Select activity Lecture 5 slides (2023): N-gram language models
Lecture 5 slides (2023): N-gram language models File PDF
lexicon
language modeling
n-grams, smoothing
Select activity Lecture 6 slides (2023): Neural network language models
Lecture 6 slides (2023): Neural network language models File PDF
Intro to NNLM
Recurrent neural network language models
Long Short-Term Memory language models
Attention
Select activity Lecture 6 NNLM exercise
Not available unless any of: You are a(n) Teacher (MC) ...
Not available unless any of:

You are a(n) Teacher (MC)

You are a(n) Teacher

You are a(n) Student
Lecture 6 NNLM exercise Assignment

Please type or upload your calculations here, e.g. as a photo, text or pdf file to earn a lecture activity point.
Select activity Lectures 7-8 - Continuous speech and decoding

Lectures 7-8 - Continuous speech and decoding
Select activity Lecture 7-8 slides (2023)
Lecture 7-8 slides (2023) File PDF
recognition in continuous speech
token passing decoder
improving the recognition performance and speed
measuring the recognition performance
Select activity Lecture 7 exercise: Token passing decoder
Not available unless any of: You are a(n) Teacher (MC) ...
Not available unless any of:

You are a(n) Teacher (MC)

You are a(n) Student

You are a(n) Teacher
Lecture 7 exercise: Token passing decoder Assignment

Fill in the last column with final probabilities of the tokens, select the best token and output the corresponding state sequence!
The goal is to verify that you have the learned the idea of the Token passing decoder. The extremely simplified HMM system is almost the same as in the 2B Viterbi algorithm exercise. The observed "sounds" are just quantified to either "A" or "B" with given probabilities in states S0 and S1. Now the task is to find the most likely state sequence that can produce the sequence of sounds A, A, B using a simple language model (LM). The toy LM used here is a look-up table that tells probabilities for different state sequences, (0,1), (0,0,1) etc., up to 3-grams.
Hint: You can either upload an edited source document, a pdf file, a photo of your notes or a text file with numbers. Whatever is easiest for you. To get the activity point the answer does not have to be correct.
Select activity Lecture 9-10 - End-to-end ASR with deep neural net...

Lecture 9-10 - End-to-end ASR with deep neural networks
Select activity Lecture 9 slides (2023)

Not available unless: You belong to any group

Lecture 9 slides (2023) File PDF
Select activity Lecture 10 slides (2023)

Not available unless: You belong to any group

Lecture 10 slides (2023) File PDF
Select activity Lecture 9-10 slides (2022)

Not available unless: You belong to any group

Lecture 9-10 slides (2022) File PDF

This is 2022, but because the content was quite different (focusing on attention-based encoder-decoder architectures) this maybe worth studying, too.
Select activity Lecture 9 exercise
Not available unless any of: You are a(n) Teacher ...
Not available unless any of:

You are a(n) Teacher

You are a(n) Student

You are a(n) Teacher (MC)
Lecture 9 exercise Assignment
Select activity Presentations for the weeks 6-7

Presentations for the weeks 6-7 File PDF

Here's the presentation schedule as finalized at the last lecture.

MyCourses service break

ELEC-E5510 - Speech Recognition D, Lecture, 25.10.2023-8.12.2023

Topic outline

Lectures

Lectures 1-2 Feature extraction and modeling

Lectures 3-4 - Phoneme modeling

Lectures 5-6 - Language Modeling

Lectures 7-8 - Continuous speech and decoding

Lecture 9-10 - End-to-end ASR with deep neural networks

Students

Teachers

About service