phy731 - Wahlpflicht Theorie (Veranstaltungsübersicht)

phy731 - Wahlpflicht Theorie (Veranstaltungsübersicht)

Institut für Physik 6 KP
Modulteile Semesterveranstaltungen Sommersemester 2023 Prüfungsleistung
Vorlesung
  • Kein Zugang 2.01.597 - Trustworthy Machine Learning Lehrende anzeigen
    • Prof. Dr. Nils Strodthoff
    • Tiezhi Wang

    Montag: 12:15 - 13:45, wöchentlich (ab 17.04.2023), Ort: V03 0-E002
    Donnerstag: 08:15 - 09:45, wöchentlich (ab 13.04.2023), Ort: V03 0-E002
    Donnerstag: 14:15 - 15:45, wöchentlich (ab 11.05.2023), Ort: V03 2-A208
    Termine am Mittwoch, 16.08.2023 09:00 - 11:15, Ort: ((V04 1-123))

    Maschinelle Lernalgorithmen finden zunehmend breite Anwendung in verschiedensten insbesondere auch sicherheitskritischen Anwendungsbereichen, doch die Qualität dieser Algorithmen wird in den seltensten Fällen systematisch untersucht. Der Schwerpunkt dieser Veranstaltung liegt auf verschiedensten Qualitätsdimensionen für maschinelle Lernalgorithmen, insbesondere tiefe neuronale Netzwerke, angefangen von der Messung der Leistungsfähigkeit, über Interpretierbarkeit/Erklärbarkeit (XAI), Robustheit (adversarial robustness, Robustheit gegen Störung im Input), Unsicherheitsquantifizierung, Distribution Shift, Domain Adaptation, Fairness/Bias bis hin zu Privacy. Die Methoden werden in der Vorlesung theoretisch eingeführt und in den Übungen praktisch implementiert und angewendet. Inhaltliche Voraussetzungen sind grundlegende theoretische Kenntnisse im Bereich des maschinellen Lernens, praktische Programmierkenntnisse in Python und im Idealfall Grundkenntnisse im Training tiefer neuronaler Netzwerke.

  • Kein Zugang 5.04.4012 - Informationsverarbeitung und Kommunikation / Information Processing and Communication Lehrende anzeigen
    • Priv.-Doz. Dr. Jörn Anemüller

    Donnerstag: 12:15 - 13:45, wöchentlich (ab 13.04.2023)
    Termine am Mittwoch, 30.08.2023 11:00 - 13:00

    Course topics: - Information processing in the brain, neurons, receptive fields - Simple classification models, the perceptron, linear discriminant analysis, regression approach to classification - Generative approaches, k-nearest neighbour classification, Bayes equation - Model selection and cross-validation - Logistic regression, binary cross-entropy loss function, gradient descent - Gradient descent optimization and regularization, multi-layer perceptron and error backpropagation - Convolutional networks, deep neural networks, receptive fields in deep netoworks - Reinforcement learning - Sequence modeling, speech recognition, markov chains, hidden markov model (HMMs) - Transformer deep networks, large language models (LLMs), from HMMs to LLMs - Information theory, measuring information, entropy - Information theory continued, entropy bound for coding, compression

  • Kein Zugang 5.04.4215 - Machine Learning II – Advanced Learning and Inference Methods Lehrende anzeigen
    • Prof. Dr. Jörg Lücke

    Donnerstag: 10:15 - 11:45, wöchentlich (ab 13.04.2023)

    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.

  • Kein Zugang 5.04.4586 - Digital Signal Processing Lehrende anzeigen
    • Prof. Dr. Simon Doclo

    Montag: 16:15 - 17:45, wöchentlich (ab 17.04.2023), Ort: W32 1-112
    Termine am Dienstag, 25.07.2023 14:15 - 15:45, Freitag, 28.07.2023, Mittwoch, 11.10.2023 10:00 - 12:00, Ort: W32 1-112, W32 0-005, W03 1-156

    Engineering Physics: Alternative für Signal- und Systemtheorie

  • Kein Zugang 5.04.4643 - Adaptive systems for speech signal processing Lehrende anzeigen
    • Prof. Dr. Gerald Enzner

    Montag: 12:15 - 13:45, wöchentlich (ab 17.04.2023)
    Montag: 14:15 - 15:45, wöchentlich (ab 17.04.2023)

    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.

Seminar
Übung
  • Kein Zugang 2.01.597 - Trustworthy Machine Learning Lehrende anzeigen
    • Prof. Dr. Nils Strodthoff
    • Tiezhi Wang

    Montag: 12:15 - 13:45, wöchentlich (ab 17.04.2023), Ort: V03 0-E002
    Donnerstag: 08:15 - 09:45, wöchentlich (ab 13.04.2023), Ort: V03 0-E002
    Donnerstag: 14:15 - 15:45, wöchentlich (ab 11.05.2023), Ort: V03 2-A208
    Termine am Mittwoch, 16.08.2023 09:00 - 11:15, Ort: ((V04 1-123))

    Maschinelle Lernalgorithmen finden zunehmend breite Anwendung in verschiedensten insbesondere auch sicherheitskritischen Anwendungsbereichen, doch die Qualität dieser Algorithmen wird in den seltensten Fällen systematisch untersucht. Der Schwerpunkt dieser Veranstaltung liegt auf verschiedensten Qualitätsdimensionen für maschinelle Lernalgorithmen, insbesondere tiefe neuronale Netzwerke, angefangen von der Messung der Leistungsfähigkeit, über Interpretierbarkeit/Erklärbarkeit (XAI), Robustheit (adversarial robustness, Robustheit gegen Störung im Input), Unsicherheitsquantifizierung, Distribution Shift, Domain Adaptation, Fairness/Bias bis hin zu Privacy. Die Methoden werden in der Vorlesung theoretisch eingeführt und in den Übungen praktisch implementiert und angewendet. Inhaltliche Voraussetzungen sind grundlegende theoretische Kenntnisse im Bereich des maschinellen Lernens, praktische Programmierkenntnisse in Python und im Idealfall Grundkenntnisse im Training tiefer neuronaler Netzwerke.

  • Kein Zugang 5.04.4012 Ü1 - Übung zu Informationsverarbeitung und Kommunikation / Information Processing and Communication Lehrende anzeigen
    • Priv.-Doz. Dr. Jörn Anemüller
    • Eike Jannik Nustede, M. Sc.

    Dienstag: 16:15 - 17:45, wöchentlich (ab 18.04.2023)

    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

  • Kein Zugang 5.04.4215 Ü1 - Exercises to Machine Learning II – Advanced Learning and Inference Methods Lehrende anzeigen
    • Prof. Dr. Jörg Lücke
    • Till Kahlke

    Dienstag: 10:15 - 11:45, wöchentlich (ab 18.04.2023), Übung

    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.

  • Kein Zugang 5.04.4215 Ü2 - Exercises to Machine Learning II – Advanced Learning and Inference Methods Lehrende anzeigen
    • Prof. Dr. Jörg Lücke
    • Filippos Panagiotou
    • Florian Hirschberger
    • Dmytro Velychko
    • Sebastian Salwig

    Dienstag: 14:15 - 15:45, wöchentlich (ab 18.04.2023), Übung

    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.

  • Kein Zugang 5.04.4586 Ü1 - Exercises to Digital Signal Processing Lehrende anzeigen
    • Henri Gode
    • Daniel Fejgin

    Mittwoch: 12:15 - 13:45, wöchentlich (ab 12.04.2023)

    Engineering Physics: Alternative für Signal- und Systemtheorie

  • Kein Zugang 5.04.4643 - Adaptive systems for speech signal processing Lehrende anzeigen
    • Prof. Dr. Gerald Enzner

    Montag: 12:15 - 13:45, wöchentlich (ab 17.04.2023)
    Montag: 14:15 - 15:45, wöchentlich (ab 17.04.2023)

    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.

Hinweise zum Modul
Teilnahmevoraussetzungen
Bachelor in Physik, Technik und Medizin oder entsprechender Abschluss
Prüfungsleistung Modul
Klausur (max. 180 Min.) oder mündliche Prüfung (30 Min.) oder Referat (30 Min.) oder Hausarbeit
Kompetenzziele
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.