Lecture "Pattern Recognition and Machine Learning"

 

Basic Information
Lecturers: Gerhard Schmidt (lecture) and Sören Lange / Julius Braun-Dullaeus (exercise)
Room: KS2/Geb.F - SR-II
E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Language: English
Target group: Students in electrical engineering and computer engineering
Prerequisites: Basics in system theory
Contents:

In this lecture the basics of speech, audio, and music signal processing are treated. Often schemes that are based on statistical optimization are utilized for these applications. The involved cost function are matched to the human audio perception.

Topic overview:

  • Preprocessing to reduce signal distortions
    • Noise reduction
    • Beamforming
  • Feature extraction and data compression
  • Pattern recognition and data regression
    • Learning types
      • Unsupervised learning
      • Supervised learning
      • Reinformcement learning
    • Codebooks
    • Object-to-vector conversion
    • Gaussian mixture models (GMMs)
    • Artificial Neural Networks (ANNs)
      • Multilayer perceptron
      • Convolutional neuronal networks (CNNs)
      • Recurrent neural networks (RNNs)
      • Autoencoder networks
      • Generative adversarial networks (GANs)
    • Hidden Markov models (HMMs)
    • Explainable artificial intelligence (EAI)

 

News

No news yet.

 

Schedule

The following schedule regarding lectures and excercises is preliminary and may be adapted during the semester. The lecture will usually take place from 14:30 h - 17:00 h. The exercise will follow after at 17:15 h - 18:00 h.

Date Lecture Exercise
20.10.2026 Introduction -
27.10.2026 Noise Suppression and Cost Functions Noise Suppression (video and exercise in the lecture room)
03.11.2026 Beamforming Beamforming (video and exercise in the lecture room)
10.11.2026 Feature Extraction Feature Extraction (video and exercise in the lecture room)
17.11.2026 Codebook Training Codebook Training (video and exercise in the lecture room)
24.11.2026 Bandwidth Extension Bandwidth Extension (video and exercise in the lecture room)
01.12.2026 Object-to-vector Conversion Object-to-vector Conversion (exercise in the lecture room)
08.12.2026 Gaussian Mixture Models (and Speaker Recognition) Gaussian Mixture Models (video and exercise in the lecture room)
15.12.2026 Neural Networks - Part 1 -
05.01.2027 Neural Networks - Part 2 Neural Networks (video and exercise in the lecture room)
12.01.2027 Student talks -
19.01.2027 Hidden Markov Models Hidden Markov Models (video and exercise in the lecture room)
26.01.2027 Explainable artificial intelligence Explainable artificial intelligence (exercise in the lecture room)
02.02.2027 TBD TBD

 

Lecture Slides

Link Content
Slides of the lecture "Introduction"
(Introduction, boundary conditions of the lecture, applications)
Slides of the lecture "Noise Suppression"
(Noise suppression, dereverberation, speech reconstruction)
Slides of the lecture "Beamforming"
(Fixed and adaptive beamforming, postfiltering)
Slides of the lecture "Feature Extraction"
(Linear prediction, cepstrum, mel-filtered cepstral coefficients)
Slides of the lecture "Codebook Training"
(K-means algorithm, LBG algorithms)
Slides of the lecture "Bandwidth Extension"
(Model-bases approaches, evaluation)
Slides of the lecture "Object-to-vector Conversion (word to vector conversion)"
(Word2Vec)
Slides of the lecture "Gaussian Mixture Models (GMMs)"
(Training with the EM algorithm, applications)
Slides of the lecture "Neural networks"
(Network types, training procecures)
Slides of the lecture "Hidden Markov Models (HMMs)"
(Efficient probability calculation, training of HMMs)
Slides of the lecture "Explainable artificial intelligence"
(Understanding what had contributed mainly to the decision of neural networks, etc.)

 

Matlab Demos

Link Content
Matlab demo (GUI based) for adaptive noise suppression
Matlab demo (GUI based) for linear prediction

 

Exercises

For each lecture topic on-demand video will be provided. There will be an in-presence exercises to discuss your questions and requested topics. You may send in questions (if you want the answer to be supported by slides) or exercise topic suggestions to This email address is being protected from spambots. You need JavaScript enabled to view it. at any time.

Video Content Material

Noise suppression:

  • Wiener Filter summary
  • comprehension questions
  • python demo

Beamforming:

  • beamforming summary
  • comprehension questions
  • python demo

Feature extraction:

  • linear prediction
  • LPC, LPCC, MFCC features
  • comprehension questions
  • python demo

Codebook training:

  • codebooks
  • codebook training
  • comprehension questions
  • python demo

Bandwidth extension:

  • bandwidth extension summary
  • comprehension questions
  • python demo

Object-to-Vector Conversion:

  • object-to-vector conversion summary
  • python demo

Gaussian Mixture Models:

  • Gaussian Mixture Models
  • EM algorithm
  • comprehension questions
  • python demo

Neural Networks:

  • neural networks summary
  • comprehension questions
  • python demo

Hidden Markov Models:

  • hidden markov models summary
  • comprehension questions
  • python demo

Speaker and Speech Recognition:

  • speaker and speech recognition summary
  • comprehension questions
  • python demo

Explainable AI:

  • comprehension questions

 

Talks

Each student will give a talk about a certain topic. The aim is both to give you the chance to work on a pattern recognition-related topic that interests you, and to improve your presentational skills. The talk is also a prerequisite for your admission to the exam. The talks should be held in English and should take 10 minutes, plus 2.5 minutes of discussion and 2.5 minutes of feedback. Please write an email to This email address is being protected from spambots. You need JavaScript enabled to view it. to reserve your topic. The talk registration deadline is not set yet.

Handouts after a presentation are a valuable addition because they help the audience retain and review the key information. During a talk, listeners may miss important details or forget specific points. A well-prepared handout summarizes the main ideas, highlights essential data, and provides useful references for further reading. These references can be especially helpful if listeners want to explore the topic in more depth and continue learning on their own. It also allows participants to focus more on listening instead of taking extensive notes. Overall, handouts support understanding, reinforce learning, and make the presentation more effective and memorable.

For the reasons mentioned above, we have prepared a template for a handout. You can find that here. Please fill (and change) it with your presentation details and send it to us (same e-mail adress as mention above) prior to your talk.

 

Evaluation

The completed evaluations of the last years can be found here. Please note that this page contains the evaluations of all of our lectures. As a consequence you must scroll down to get to the evaluations of this lecture.

 

Exams

Remember to register for the exam in the QiS system! And book a time-slot for your oral exam using this form.