The students gain a broad operational perspective for the design of speech adaptive systems and respective algorithms with a particular focus on adaptive digital filters. The important NLMS, RLS, FDAF and Kalman-Filter algorithms can be derived from fundamental principles. Diverse applications from speech and acoustic signal processing deliver practical insight into the utilization of the fundamentals, for instance, in acoustic noise reduction, echo cancellation, dereverberation, acoustic channel estimation and equalization. However, the acquired knowledge allows for a broader interpretation in the context of engineering and physics. The computer exercises of larger scale will teach the students to argue, select and evaluate algorithms for the problem at hand. By discussion in the panel, students learn to demonstrate, defend and trade their solution against others. Theoretical exercises finally deliver the ability to argue and prove a speech processing design with the appropriate vocabulary.
Die Studierenden erlernen, wie statistische Eigenschaften von Signalen zur Lösung von Problemen der Angewandten Physik, insbesondere der Klassifikation, parametrischen Modellierung und Übertragung von Signalen genutzt werden können. Theoretische Lernziele beinhalten damit eine Wiederholung und Festigung statistischer Grundlagen und eine Verständnis von deren Nutzung für Algorithmen unterschiedlicher Zielsetzung und Komplexität. Im praktischen Teil werden Eigenschaften der behandelten Methoden selbständig erarbeitet sowie Algorithmen auf dem Rechner implementiert und auf reale Daten angewendet, so daß der Umgang mit theoretischen Konzepten und ihre praktische Umsetzung erlernt werden.
Inhalte:
Grundfragen der Informationsverarbeitung (Klassifikation, Regression, Clustering), Lösungsmethoden basierend auf Dichteschätzung und diskriminativen Ansätzen (z.B. Bayes Schätzung, k-nearest neighbour, Hauptkomponentenanalyse, support-vector-machines, Hidden-Markov- Modelle), Grundlagen der Informationstheorie, Methoden der analogen und digitalen Nachrichtenübertragung, Prinzipien der Kanalcodierung und Kompression
Machine learning algorithms find its way into an increasing number of
(safety-critical) application domains but their quality is rarely assessed in a systematic way. The focus of this module are quality criteria for machine learning algorithms, in particular of deep neural networks, ranging from performance evaluation over explainability/interpretability (XAI), robustness (adversarial robustness, robustness against input perturbations), uncertainty quantification, distribution shift, domain adaptation, fairness/bias to privacy. The methods are introduced theoretically in the lecture and implemented/applied practically in the exercises. Prerequisites are basic theoretical knowledge in machine learning, programming skills in Python and ideally practical knowledge in training neural networks.
The students will deepen their knowledge on mathematical models of data and sensory signals. Building up on the previously acquired Machine Learning models and methods, the students will be lead closer to current research topics and will learn about models that currently represent the state-of-the-art. Based on these models, the students will be exposed to the typical theoretical and practical challenges in the development of current Machine Learning algorithms. Typical such challenges are analytical and computational intractabilities, or local optima problems. Based on concrete examples, the students will learn how to address such problems. Applications to different data will teach skills to use the appropriate model for a desired task and the ability to interpret an algorithm’s result as well as ways for further improvements. Furthermore, the students will learn interpretations of biological and artificial intelligence based on state-of-the-art Machine Learning models.
Contents:
This course builds up on the basic models and methods introduced in introductory Machine Learning lectures. Advanced Machine Learning models will be introduced alongside methods for efficient parameter optimization. Analytical approximations for computationally intractable models will be defined and discussed as well as stochastic (Monte Carlo) approximations. Advantages of different approximations will be contrasted with their potential disadvantages. Advanced models in the lecture will include models for clustering, classification, recognition, denoising, compression, dimensionality reduction, deep learning, tracking etc. Typical application domains will be general pattern recognition, computational neuroscience and sensory data models including computer hearing and computer vision.
The students gain a broad operational perspective for the design of speech adaptive systems and respective algorithms with a particular focus on adaptive digital filters. The important NLMS, RLS, FDAF and Kalman-Filter algorithms can be derived from fundamental principles. Diverse applications from speech and acoustic signal processing deliver practical insight into the utilization of the fundamentals, for instance, in acoustic noise reduction, echo cancellation, dereverberation, acoustic channel estimation and equalization. However, the acquired knowledge allows for a broader interpretation in the context of engineering and physics. The computer exercises of larger scale will teach the students to argue, select and evaluate algorithms for the problem at hand. By discussion in the panel, students learn to demonstrate, defend and trade their solution against others. Theoretical exercises finally deliver the ability to argue and prove a speech processing design with the appropriate vocabulary.
Die Studierenden erlernen, wie statistische Eigenschaften von Signalen zur Lösung von Problemen der Angewandten Physik, insbesondere der Klassifikation, parametrischen Modellierung und Übertragung von Signalen genutzt werden können. Theoretische Lernziele beinhalten damit eine Wiederholung und Festigung statistischer Grundlagen und eine Verständnis von deren Nutzung für Algorithmen unterschiedlicher Zielsetzung und Komplexität. Im praktischen Teil werden Eigenschaften der behandelten Methoden selbständig erarbeitet sowie Algorithmen auf dem Rechner implementiert und auf reale Daten angewendet, so daß der Umgang mit theoretischen Konzepten und ihre praktische Umsetzung erlernt werden.
Inhalte:
Grundfragen der Informationsverarbeitung (Klassifikation, Regression, Clustering), Lösungsmethoden basierend auf Dichteschätzung und diskriminativen Ansätzen (z.B. Bayes Schätzung, k-nearest neighbour, Hauptkomponentenanalyse, support-vector-machines, Hidden-Markov- Modelle), Grundlagen der Informationstheorie, Methoden der analogen und digitalen Nachrichtenübertragung, Prinzipien der Kanalcodierung und Kompression
The students will deepen their knowledge on mathematical models of data and sensory signals. Building up on the previously acquired Machine Learning models and methods, the students will be lead closer to current research topics and will learn about models that currently represent the state-of-the-art. Based on these models, the students will be exposed to the typical theoretical and practical challenges in the development of current Machine Learning algorithms. Typical such challenges are analytical and computational intractabilities, or local optima problems. Based on concrete examples, the students will learn how to address such problems. Applications to different data will teach skills to use the appropriate model for a desired task and the ability to interpret an algorithm’s result as well as ways for further improvements. Furthermore, the students will learn interpretations of biological and artificial intelligence based on state-of-the-art Machine Learning models.
Contents:
This course builds up on the basic models and methods introduced in introductory Machine Learning lectures. Advanced Machine Learning models will be introduced alongside methods for efficient parameter optimization. Analytical approximations for computationally intractable models will be defined and discussed as well as stochastic (Monte Carlo) approximations. Advantages of different approximations will be contrasted with their potential disadvantages. Advanced models in the lecture will include models for clustering, classification, recognition, denoising, compression, dimensionality reduction, deep learning, tracking etc. Typical application domains will be general pattern recognition, computational neuroscience and sensory data models including computer hearing and computer vision.
Machine learning algorithms find its way into an increasing number of
(safety-critical) application domains but their quality is rarely assessed in a systematic way. The focus of this module are quality criteria for machine learning algorithms, in particular of deep neural networks, ranging from performance evaluation over explainability/interpretability (XAI), robustness (adversarial robustness, robustness against input perturbations), uncertainty quantification, distribution shift, domain adaptation, fairness/bias to privacy. The methods are introduced theoretically in the lecture and implemented/applied practically in the exercises. Prerequisites are basic theoretical knowledge in machine learning, programming skills in Python and ideally practical knowledge in training neural networks.
The students will deepen their knowledge on mathematical models of data and sensory signals. Building up on the previously acquired Machine Learning models and methods, the students will be lead closer to current research topics and will learn about models that currently represent the state-of-the-art. Based on these models, the students will be exposed to the typical theoretical and practical challenges in the development of current Machine Learning algorithms. Typical such challenges are analytical and computational intractabilities, or local optima problems. Based on concrete examples, the students will learn how to address such problems. Applications to different data will teach skills to use the appropriate model for a desired task and the ability to interpret an algorithm’s result as well as ways for further improvements. Furthermore, the students will learn interpretations of biological and artificial intelligence based on state-of-the-art Machine Learning models.
Contents:
This course builds up on the basic models and methods introduced in introductory Machine Learning lectures. Advanced Machine Learning models will be introduced alongside methods for efficient parameter optimization. Analytical approximations for computationally intractable models will be defined and discussed as well as stochastic (Monte Carlo) approximations. Advantages of different approximations will be contrasted with their potential disadvantages. Advanced models in the lecture will include models for clustering, classification, recognition, denoising, compression, dimensionality reduction, deep learning, tracking etc. Typical application domains will be general pattern recognition, computational neuroscience and sensory data models including computer hearing and computer vision.
Engineering Physics: Alternative für Signal- und Systemtheorie
Hinweise zum Modul
Prerequisites
Bachelor in Physik, Technik und Medizin oder entsprechender Abschluss
Module examination
M
Skills to be acquired in this module
Die Studierenden erwerben die theoretischen Voraussetzungen für die numerische und analytische Modellierung komplexer Vorgänge in der Medizin, Biologie und Biophysik, und wenden Forschungsmethoden des Exzellenzcluster Hearing4all im Modellierungsbereich an. Spezielle Kompetenzen abhängig von der gewählten Veranstaltung.