Student Topics and Bachelor's or Master's Theses

Regarding topics for Bachelor's theses, projects, or Master's theses, if you are interested in working, e.g., on

• Boolean satisfiability (SAT) or answer set programming (ASP),
• declarative problem solving (e.g., solving problems using SAT or ASP), 
• efficient algorithms for challenging (NP-hard) problems, 
• logic & computation, logical languages, 
• knowledge representation & reasoning,
• computational complexity of reasoning,
• non-monotonic reasoning, or
• preferential reasoning and computational social choice, 

you can contact Johannes P. Wallner for more information for student topics at Graz University of Technology. You can also have a look at the group's webpage and recent publications (see links on the right side), to get a feeling on potential topics. 

A thesis can be oriented more theoretically, or more on the implementation side. The concrete topic is discussed with the supervisor.

Here is a list of potential topics for a project or thesis. Note that the concrete topics to be worked on are discussed in-person. This list is not always up-to-date.

1. Learning of strict and defeasible rules
   A strict rule would be similar to a logical implication A -> B (if A then necessarily B). A defeasible rule is closer to a default (if A then by default B), however with exceptions. Here the direction is to work on how to learn such rules. Basis could be works in inductive logic programming (ILP).
2. Implementing, e.g.,  via SAT or ASP based methods, different variants of closed-world reasoning. 
   Closed world assumption, simply put, states that what is not explicitly mentioned is assumed false. This can be interpreted in several ways. This direction is to explore implementations for various ways of interpreting closed world assumptions.
3. Implement preferred subtheories.
   Logical reasoning, classically in, e.g., propositional logic, amounts to infer a deductive conclusion form a given formula. In case of an inconsistent (unsatisfiable) formula no useful deductive conclusions can be derived (classically). In such a case, maximal parts of the given formula(s) can be used to get maximal consistent subsets, for reasoning. Preferred subtheories augment these with preferential reasoning (preferring some formulas over others). This direction explores implementing reasoning for preferred subtheories, likely utilizing SAT, MaxSAT, or ASP.
4. Investigating possiblities for implementing logics interpreted under team semantics.
   Team semantics are a different way to interpret (propositional) logic sometimes aimed to incorporate plurality of data. This direction explores possible ways of implementing team-based semantics of propositional logic, and possible reasoning tasks to be implemented.
5. Absorptive semirings for obtaining small derivations.
   Semirings have been studied in various aspects, e.g., connected to provenance (e.g., to answer from which sources a query answer was derived). Some semirings allow for certain optimizations of derivations. This direction explores possibilities of utilizing such semirings for optimizing derivations.
6. Logical expressivity of certain machine learning architectures.
   Recently, some machine learning architectures, e.g., graph neural networks, have been connected to properties expressible in certain logics. This direction explores expressivity of these machine learning architectures. 
7. Formalize selected proofs in a formal language for proof assistants.
   Nowadays, proof assistants, such as first-order logic theorem provers (e.g., vampire), higher-order logic (e.g., also Isabelle), or lean, offer ways of checking or semi-automating proofs. This direction explores ways of formalizing selected proofs in formal languages of such proof assistants.

2025/26 (planned):

• winter term:
  • Logic-based Knowledge Representation (VU)
  • seminars
• summer term:
  • Artificial Intelligence 2 (VU)
  • Declarative Programming (VU)
  • Intelligent Systems (VO and KU)

2024/25:

• winter term: 
  • Foundations in Computer Science (several courses in TU Graz Online)
  • Logic-based Knowledge Representation (VU)
• summer term: 
  • Basics in Artificial Intelligence and Logic (VU)
  • Intelligent Systems (VO and KU)

2023/24:

• summer term: Basics in Artificial Intelligence and Logic (at TU Graz Online)
• winter term: Logic-based Knowledge Representation (at TU Graz Online)
• temporary replacement for Grundlagen der Informatik at FH Campus02 and FH Joanneum

2022/23:

• summer term: Basics in Artificial Intelligence and Logic (at TU Graz Online)
• News course winter term: Logic-based Knowledge Representation (at TU Graz Online)
• temporary replacement for Grundlagen der Informatik at FH Campus02 and FH Joanneum

2021/22:

• Basics in Artificial Intelligence and Logic (at TU Graz Online)

2020 and before: teaching at TU Wien and University of Helsinki