Manuel Milling

Manuel Milling

he/him

Deep Learning Researcher and Teacher

Teaching

Below is a comprehensive list of teaching efforts during my PhD and post-doc time, which have been taken place at University of Augsburg (UA) until 2023 and at TUM since 2024 in the respective TUM Schools of Computation, Information and Technology (CIT) and Medicine and Health (MH), as well as UA Faculty of Applied Computer Science (FAI). Since 2019 I have taken the lead in coordinating and co-designing teaching activities in the respective groups. Between May 2022 and November 2023 I served as recognition officer to process recognition requests from students in the bachelor and master program of medical informatics.

Structured Courses

The core target audience is highlighted, even though related study programs are eligible as well.

Advanced Practical Deep Learning for Audio, Speech, and Language (TUM CIT)

10 ECTS, Master-level practical, Informatics, group projects and regular exchange formats (SS25, WS25/26, SS26)
Co-lead design of course and material, teaching

Healthcare Innovation Program (TUM Interdisciplinary)

5 ECTS, Master-level lecture, Informatics, Medicine and Business, group projects and regular exchange formats (SS26)
Co-lead design and held a guest lecture, contributed to course organisation and student mentoring

Introduction to Programming and Artificial Intelligence in Sports Science (TUM MH)

5 ECTS, Bachelor-level lecture-style seminar and tutorial, Sports Science (WS25/26)
Lead design of course and material (lectures and exercises), teaching (lectures and exercises)

Deep Learning (UA FAI)

5 ECTS, Master-level lecture and tutorial, Informatics (SS19, WS1920, SS22, SS23)
Lead design of course and material (exercises), teaching (exercises and occasionally lectures)

Practical Deep Learning (UA FAI)

5 ECTS, Bachelor-level practical, Informatics (SS19, SS20, SS21, SS22, WS22/23, SS23)
Lead design of course and material (lectures and exercises), teaching (lectures and exercises)

Practical Speech Pathology (UA FAI)

5 ECTS, Bachelor-level practical, Informatics (WS19/20)
Contribute to design of course material (exercises), contribute to teaching (exercises)

Thesis Supervision

Master Theses

  • Sharpness as a Scoring Function in Curriculum Learning (2026)
  • Design and Implementation of the Spezi Study Platform for Patient‑Generated Health Data, co-supervised with Stanford Biodesign (2026)
  • Designing a Personalized, Explainable Feedback Pipeline using Smartphone-Based Behavioral Data, co-supervised with LMU Department of Psychology (2026)
  • Improving Instruction Adherence in Long-Form Generation via Parallel Writer-Critic Loops, co-supervised with Forgent AI (2026)
  • Development of a Language Model for the German Medical Domain, co-supervised with TUM Chair of Medical Informatics (2025)
  • Building a Digital Health Development Ecosystem for Native Mobile Applications, co-supervised with Stanford Biodesign (2024)
  • Towards Predictive Maintenance: Deep Learning-Based Prediction of Battery Failures in Trucks, co-supervised with MAN Truck and Bus (2024)
  • Insights from Curriculum Learning Across Modalities in a Deep Learning Setting (2024)
  • Airborne Pollen Grain Detection from Partially Labelled Data utilising Semisupervised Learning (2022)
  • Deep Learning Drives Mobile Application: Listen to Image-Based Audio (2022)
  • Speaker Diarisation in Therapeutic Sessions for Children with ASD: An Evaluation of Supervised Deep Learning-Based Approaches (2021)
  • Adapting Deep Learning Classifiers for Online Personalisation - Robust Gesture and Activity Recognition on Consumer Devices, co-supervised with Huawei (2021)
  • Automated Detection and Classification of Airborne Pollen Grains Using Deep Learning (2020)

Bachelor Theses

  • Dual-Contrastive Sparse Autoencoders Reveal Features of Musical Interpretation, co-supervised with MIT Media Lab (2026)
  • Development of an Affect-Detection System and Implementation in the Phone Study App, co-supervised with LMU Department of Psychology (2026)
  • Enhancing Wearable Human Activity Recognition through Smartphone and Smartwatch Sensor Fusion, co-supervised with ETH Zürich Information Management (2025)
  • A Platform for Evaluating Large Language Models for Personalized Medical Consultations, co-supervised with Stanford Biodesign (2025)
  • Interpretable Deep Learning for Dysarthria Detection from Speech (2025)
  • Earth Mover Distance-Guided learning for Age Prediction from Speech (2023)
  • Altersvorhersage auf Bildern: Deep Learning Klassifizierer mit Distanzbasierten Verlustfunktionen (2023)
  • Applying Recent CNN Architectures for Automated Classification of Airborne Pollen Grains (2023)
  • Machine Learning-based Prediction of PDDS and PHQ-8 Scores in Persons with Multiple Sclerosis from Passive Audio Data (2022)
  • Beyond First-Order Optimisation for Deep Learning-Based Audio Tasks (2021)
  • Deep Learning-Based Localisation of Adherent Cells on Microscope Images (2021)
  • Deep Learning-Based Speech Emotion Recognition: A Cross-Cultural Analysis for Children with Autistic Spectrum Condition (2021)
  • Deep Learning-Based Price Prediction for Online Game Trade Markets (2020)
  • Detecting and Tracking the Migration of Adherent Cells Influenced by Surface Acoustic Waves with Deep Convolutional Neural Networks (2020)

Clinical Application Projects

CAPs are bachelor-level guided research projects for TUM CS students with a minor in medicine with a volume of 6 ECTS that are this case centered around topics of ML and data science on medical data.

  • Analysis of Vital Signs for Smoking Recognition (2026)
  • Exploring Deep Learning Methodologies for Depression Detection (2025)
  • Recognizing ADHD from EEG signals with the help of machine and deep learning (2025)
  • Investigating Respiratory Sounds with Machine and Deep Learning (2025)
  • Investigating Convolutional Neural Networks for Brain Tumor Recognition from MRI Images (2025)

Seminar Theses

The following list concludes seminar theses written by informatics students in the context of the seminars Embedded Intelligence for Health Care and Wellbeing, Computer Audition, Sports Informatics, Computational Intelligence, and Digital Health

  • Technology Adoption for Children with Neurodevelopmental Disorders (2023)
  • Recent Trends in Applications of Automatic Emotion Recognition for Autistic Children (2023)
  • Understanding Data Augmentation (2023)
  • Sharpness Measures for Loss Function Minima (2023)
  • Acoustic Feature Analysis of Child Voice Activity: Typical Development vs. Autism Spectrum Disorder (2022)
  • Emotion recognition for autistic individuals: Datasets, Approaches (2022)
  • Speech-based Emotion Conversion (2021)
  • Unsupervised object Detection and Segmentation with Deep Learning: A Review (2020)
  • Score Prediction through Tennis Grunting using Deep Learning Methods (2020)
  • War das ein Punkt? Analyse der Laute im Tennis (2020)
  • Similarity in the context of neural networks (2020)
  • Recent Trends for Optimisation Algorithms in Deep Learning (2020)
  • LARS and LAMB in comparison to some other optimisation algorithms (2020)
  • Measurement of Rapport Over Time (2019)
  • Audio-Based Classification of Sports (2019)

Other supervised projects

  • Implementation of data augmentation functionality for autrainer, Project Module (2024)
  • Exploring Distance-based Loss Functions for Deep Learning in Classification Tasks, Project Module (2020)
  • Deep Speaker Identification, Project Module (2020)
  • Narcissism detection from text, Project Module (2020)
  • A Survey on Neural Network Optimization for Audio Classification Tasks, Research Module (2020)