Vorlesungen

Vorlesung/Lecture:
Do 12 – 14 Uhr c.t., Geschwister-Scholl-Platz 1, Hauptgebäude, M 105
16.04.-16.07.2026
2 SWS

Übung/Exercises:
Fr 10 – 12 Uhr c.t., Geschwister-Scholl.Platz 1, Hauptgebäude, D 209
17.04.-17.07.2026
2 SWS

This lecture focuses on Generative AI, deep learning approaches in computer vision that are generative so they cannot only analyze existing scenes, but in particular synthesize novel images and video.
Modern deep learning has fundamentally changed artificial intelligence. Computer vision was at the forefront of many of these developments and has tremendously benefited over the last decade from this progress. Novel applications as well as significant improvements to old problems continue to appear at a staggering rate. Especially the areas of image and video synthesis and understanding have seen previously unthinkable improvements and provided astounding visual results with wide-ranging implications (trustworthiness of AI, deep fakes).
We will discuss how a computer can learn to understand images and videos based on deep neural networks. The lecture will briefly review the necessary foundations of deep learning and computer vision and then cover the latest works from the quickly developing field of Generative AI. The practical exercises that accompany this course will provide hands-on experience and allow attendees to practice while building and experimenting with powerful image generation architectures.
Topics include but are not limited to:
* Image & video synthesis
* Visual superresolution and Image completion
* Artistic style transfer
* Interpretability, trustworthyness of deep models
* Self-supervised learning
* Modern deep learning approaches, such as transformers and self-attention, invertible neural networks, diffusion models, flow matching etc.

This master-level lecture surveys the core implementation techniques behind modern database management systems, with a strong systems-oriented focus. Topics include on-disk and in-memory storage layouts, buffer pool management, indexing structures (B-trees, extendible and linear hashing), query execution operators, and cost-based optimization. The course also covers transaction processing, concurrency control protocols, write-ahead logging, crash recovery, and the fundamentals of distributed query processing and replication. Through readings, design discussions, and programming projects, students analyze real system architectures and learn how theoretical ideas translate into high-performance, fault-tolerant database engines.

• Lecture: Wednesday 8-10, U127, Oettingenstr. 67
• Lab: Wednesday 12-14, U127, Oettingenstr. 67
• Enrollment key: AtomicFenceSS26

• Memory consistency models
• C++ atomic types and operations
• Synchronization algorithms (barriers, locks)
• Concurrent objects
• Lock-freedom and wait-freedom

This lecture brings together two trending topics of technology-enhanced learning that have much overlap in terms of providing feedback and self-reflection.

This course introduces the proof assistant Lean 4, its type-theoretic foundations, and its applications to computer science and mathematics.

Proof assistants are pieces of software that can be used to check the correctness of a specification of a program or the proof of a mathematical theorem. In the practical work, we learn to use Lean. We will see how to use the system to prove mathematical theorems in a precise, formal way, and how to verify small functional programs. In the course, we focus on Lean's dependent type theory and on the Curry–Howard correspondence between proofs and functional programs (λ-terms). These concepts are the basis of Lean but also of other popular systems, including Agda, Matita, and Rocq.

There are no formal prerequisites, but familiarity with functional programming (e.g., Haskell) and basic algebra is an asset. If you are new to functional programming, we recommend that you read the first chapters of Learn You a Haskell for Great Good!, stopping at the section "Only folds and horses."

Overview

Voting is an important part of democratic societies and potentially has a broad impact. Yet, with or without the use of modern technology, voting is full of algorithmic and security challenges, and the failure to address these challenges in a controlled manner may produce fundamental flaws in the voting system and potentially undermine critical societal aspects.

In this lecture, we discuss voting systems from various perspectives, notably social choice theory, security, and cryptography. What should a voting system fulfill? When is a voting system secure, even independent of the involved software? Which mechanisms should be investigated for that matter? Which methods are suitable to address these challenges?

We will investigate cryptographic voting systems, algorithmic tallying procedures, statistical methods to test the reliability of an election result, and distinguish the different layers of a voting system.

Preliminary Knowledge

Foundations on cryptography and security, as taught, for instance, in the lecture "IT-Sicherheit", are recommended. Moreover, we make an effort to coordinate the cryptography parts of the lecture with the practical "Cryptography" that you can optionally take in parallel.

Advanced Analytics and Machine Learning examines the algorithmic and systems foundations required to realize modern machine learning systems. As many workloads increasingly rely on large datasets and foundation models, performance and reliability are determined not only by learning algorithms, but also by the underlying infrastructure for data management, distributed computation, and operational deployment. The course therefore treats machine learning as an end to end pipeline, spanning data ingestion and storage, large scale processing, model training and evaluation, and inference under resource and latency constraints.

The course develops a principled understanding of scalability through core concepts in high performance computing  and distributed systems, including data locality, communication costs, synchronization, fault tolerance, and scheduling. These principles are connected to practical implementations in widely used platforms for batch and streaming analytics (e.g., Spark, Dask, Ray, Flink, Kafka) and deep learning toolchains (PyTorch and the Transformers ecosystem). Particular attention is given to infrastructure and system issues. The curriculum is complemented by responsible AI considerations and an overview of quantum machine learning as an emerging technology.

Topics: 

• Foundations of scalable analytics and ML systems (parallelism models, distributed abstractions, communication and synchronization costs, fault tolerance, scheduling)
• Data management and large scale processing (data lake architectures, SQL on semi structured data with Hive, Spark SQL, Presto, batch processing with Spark, Dask, Ray)
• Streaming systems and continuous analytics (stream processing with Kafka and Spark Streaming, state, windows, operational semantics)
• Training systems for machine learning (classical ML workflows at scale, deep learning with PyTorch, distributed training principles, checkpointing and performance analysis)
• Inference systems and operational ML (deployment patterns, throughput and latency modeling, efficiency techniques such as batching and quantization basics, monitoring concepts) 
• Emerging technologies (e.g., quantum machine learning).

Dates: 

• Saturday, March 7, 2026 
• Saturday, March 14, 2026 
• Saturday, March 21, 2026
• Saturday, March 28, 2026
• Saturday, April 18, 2026.