In dieser Arbeitsgemeinschaft (AG) werden anspruchsvolle Themen der Künstlichen Intelligenz und des Machine Learning in enger Zusammenarbeit mit ausgewählten Studenten behandelt u.a. im Rahmen von Projekt- und Abschlussarbeiten.

In den Sitzungen der AG präsentieren und diskutieren wissenschaftliche Mitarbeiter des Lehrstuhls ihren aktuellen Forschungsstand (Progress Report) und wichtige Veröffentlichungen (Journal Club) im Bereich der Künstlichen Intelligenz.

Computer Games and Games related formats are an essential branch of the media industry, with sales exceeding those of the music or the movie industry. In many games, building a dynamic environment with autonomously acting entities is necessary. This comprises any types of mobile objects, non-player characters, computer opponents, or the dynamics of the environment itself. To model these elements, techniques from the area of Artificial Intelligence allow for modeling adaptive environments with interesting dynamics. From the point of view of AI Research, games currently provide multiple environments that allow the development of breakthrough technology in Artificial Intelligence and Deep Learning. Projects like OpenAIGym, AlphaGo, OpenAI5, and Alpha-Star earned much attention in the AI research community and the broad public. The reason for the importance of games for developing autonomous systems is that games provide environments that usually allow fast throughputs and provide clearly defined tasks for a learning agent to accomplish. The lecture provides an overview of techniques for building environment engines and making these suitable for large-scale, high-throughput games and simulations.
Furthermore, we will discuss the foundations of modeling agent behavior and how to evaluate it in deterministic and non-deterministic settings. Based on this formalism, we will discuss how to analyze and predict agent or player behavior. Finally, we will introduce various techniques for optimizing agent behavior, such as sequential planning and reinforcement learning.

Enrollment Key: ReadyPlayer#1

The enrolment key can be found on the course website: https://www.nm.ifi.lmu.de/teaching/Vorlesungen/2024ss/quantum-computing/

• Lecture Date/Location: Thursdays, 12:15 – 13:45 / Geschwister-Scholl-Platz 1, Main Building, Room E 004
  
• Excercises Date/Location: Fridays, 10:15 – 11:45 / Geschwister-Scholl-Platz 1, Main Building, Room E 004

Self enrollment key: GenAISynth
Febr 19 - Apr 15

Einschreibeschlüssel/registration key: MSE

This lecture will be in held English.

Target audience: master level, advanced bachelor level (Wahlpflicht).

There are no hard requirements but we will assume you have some knowledge and experience in object-oriented programming from lectures such as introduction to programming and computer science, software engineering, and the practical programming courses.

The topic of "preferences" has attracted considerable attention in Artificial Intelligence (AI) research in the recent past, notably in the emerging field of generative AI and large language models. Preferences provide a means for specifying desires in a declarative way, which is a point of critical importance for AI. Drawing on past research on knowledge representation and reasoning, AI offers qualitative and symbolic methods for treating preferences that can reasonably complement hitherto existing approaches from other fields, such as decision theory. Needless to say, however, the acquisition of preference information is not always an easy task. Therefore, not only are modeling languages and suitable representation formalisms needed, but also methods for the automatic learning, discovery, and adaptation of preferences. This has initiated the field of preference learning, which has established itself as an independent branch of machine learning in the last decade. Broadly speaking, preference learning is concerned with methods for learning preference models from explicit or implicit preference information, which are typically used for predicting the preferences of an individual or a group of individuals in new decision contexts. While research on preference learning has been specifically triggered by applications such as learning to rank (e.g., for information retrieval), recommender systems, and reinforcement learning from human feedback (RLHF), the methods developed in this field are useful in many other domains as well.

Please enrol yourself with the following key: PLR24

Content

The course targets first-year PhD students and last-year Master students, i.e., if you have just started as wissenschaftlicher Mitarbeiter or if you study in software engineering (projects or Master thesis) and want to learn more about topics in academic work environments. The idea is to cover things that nobody told you yet in the computer-science studies but which you need if you want to go for PhD.

In order to get started with the discussion, we will usually start presenting some material. This will automatically turn into a discussion after a while. If you are interested, please sign up and stop by.

Outline

• Sciences: Wissen und Wissenschaft, Wissenschaft als Wissenskompression
• Publications: Open Access and Open Science, Repositories, Bibliographic Data, Licenses
• Artifacts
• Bibliographic Data and Rankings, Impact of Rankings
• Clustering of Research Areas
• Peer Reviews
• Authorship Criteria
• Conflict of Interest
• Plagiarism
• Evaluation of Research Results, Competitions, Proofs, Experiments
• Technical Writing, esp. with LaTeX
• Structure of Research Papers
• Grant Proposals
• Team Work, Groups, Leadership
• Principles

Please note: This course is mostly about how to do scientific work in the area of software engineering, it is not a course about software engineering itself. If you are interested in an advanced software-engineering course, we recommend Methods in Software Engineering.

Organization

There is no exam for this course, and subsequently no ECTS points. We can provide a participant certificate if you regularly attended the course.

The course happens weekly on Tuesday 14-16.

Registration key: SPSE24

Attendees

Doctoral researchers and Master students interested in a PhD.

This course covers advanced techniques for automatic software verification, especially those in the field of software model checking. It continues the Bachelor course Formal Verification and Specification (FSV). Knowledge from FSV is helpful but not mandatory. This course can be used for the specialization "Programming, Software Verification, and Logic" in the MSc computer science (cf. German site on specializations).

Topics

The course will cover the following topics:

1. Mathematical foundation for software verification
2. Configurable program analysis
3. Strongest postcondition
4. Predicate abstraction with a fixed precision
5. Craig interpolation and abstraction refinement (CEGAR)
6. Predicate abstraction with precision adjustment
7. Bounded model checking and k-induction
8. Observer automata
9. Verification witnesses
10. Test generation and symbolic execution
11. Cooperative verification
12. Project: implementing a software verifier (2 weeks)

Reference Materials

• Combining Model Checking and Data-Flow Analysis (Chapter 16 in the Handbook of Model Checking)
• A Unifying View on SMT-Based Software Verification

Organization

The course consists of weekly lectures and tutorials. Important announcements are sent via Moodle messages.

Time Slots and Rooms

• Lecture: 08:15 - 09:45, Wednesday, in Room 061, Oettingenstr. 67 (by Prof. Dr. Dirk Beyer)
• Tutorial: 14:15 - 15:45, Thursday, in C 022, HGB (by Marek Jankola and Niann-Tzer Li, Ph.D.)

The first lecture session on 2024-04-17 will be about the detailed course organization and expectations. The first tutorial session will be on 2024-04-18.

Enrolment

Please enroll yourself with the key: t1/F0al,KdUzqgbf6-JW