The color coding means:

: all participate

: Doctoral students participate

: next seminar

Please find the abstracts below the schedule and note that some of the slides are made available here, internally.

Date | Time | Topic | Speaker | Room |
---|---|---|---|---|

2019 April 2 | 13:00-16:00 | Probabilistic Reasoning with Conflicting Information | Nico Potyka (Universität Osnabrück) |
TU Dresden APB E005 |

2019 April 9 | 12:45‑15:45 | Introduction to ω-Algebraic Systems Or: Have you ever stayed awake at night because you wondered what quemirings are? |
Sven Dziadek (Universität Leipzig) |
Universität Leipzig P 501 |

Variations of nondeterminism using the example of alternating automata Or: Have you ever stayed awake at night because you wondered what nondeterminism is? |
Gustav Grabolle |
|||

2019 April 16 | 12:45‑15:45 | Answer Set Programming in Time | Torsten Schaub (Uni Potsdam) |
Universität Leipzig P 501 |

2019 April 30 | 13:00-16:00 | Argument-based paraconsistent logics | Leila Amgoud (CNRS, Toulouse) |
TU Dresden APB E005 |

2019 May 7 | 12:45‑15:45 | The Abstract State Machines Method for High Level System Design and Analysis | Egon Börger (Uni Pisa) |
Universität Leipzig P 501 |

2019 May 14 | 13:00-16:00 | TU Dresden APB E005 |
||

2019 May 28 | 13:00-16:00 | TBA | Mateus de Oliveira Oliveira (University of Bergen) |
TU Dresden APB E005 |

2019 June 4 | 12:45‑15:45 | Universität Leipzig P 501 |
||

2019 June 18 | 12:45‑15:45 | TBA | Frank Drewes (Umeå University) |
Universität Leipzig P 501 |

2019 June 25 | 13:00-16:00 | TBA |
Rafael Penaloza |
TU Dresden APB E005 |

2019 July 2 | 12:45‑15:45 | TU Dresden APB E005 |
||

2019 July 9 | 13:00-16:00 | TBA | Stefan Göller (Universität Kassel) |
Universität Leipzig P 501 |

### Abstracts

### Nico Potyka: *Probabilistic Reasoning with Conflicting Information*

Probabilistic logics generalize classical logics by extending the set of classical truth values {0, 1} to the probability interval [0, 1]. In this context, a formula does no longer necessarily evaluate to true or false, but is true with a certain probability. Probability 1 corresponds to classical truth and probability 0 to classical falsity.

In practice, probabilistic knowledge bases often contain conflicting information. For example, empirical probabilities from different statistical surveys should show a similar trend, but are unlikely to be consistent when point probabilities are considered. Similarly, when using subjective probabilities, it is unlikely that two different experts will give completely consistent judgements about the likelihood of events. Since the human brain seems not particularly well suited to grasp probabilities, even probability statements made by a single expert can be inconsistent. However, often these inconsistencies are not substantial and by relaxing probabilistic constraints slightly, we can still derive interesting information that respects the empirical evidence or the experts' intuition "as far as possible".

In this talk, I will first give a quick overview of a simple probabilistic reasoning formalism and some basic approaches that have been considered in order to deal with inconsistencies. Afterwards, we will look at one particular approach to reason over inconsistent probabilistic knowledge bases in more detail. I will explain how to deal with flat and hierarchical inconsistent knowledge bases. Interestingly, the inconsistency-tolerant reasoning problems are often „not significantly harder“ than their classical probabilistic counterparts. They also have some interesting analytical guarantees. For example, if a knowledge base is "almost consistent", the derived probabilities will be "almost classical probabilistic".

### Sven Dziadek: *Introduction to ω-Algebraic Systems*

Or: Have you ever stayed awake at night because you wondered what quemirings are?

Or: Have you ever stayed awake at night because you wondered what quemirings are?

The talk will give an introduction to algebraic systems for finite and infinite words and introduce their algebraic structures. Additionally, we give an idea how to construct the Greibach normal form for ω-algebraic systems.

### Gustav Grabolle: *Variations of nondeterminism using the example of alternating automata*

Or: Have you ever stayed awake at night because you wondered what nondeterminism is?

Or: Have you ever stayed awake at night because you wondered what nondeterminism is?

This talk will give an introduction to alternating automata of both kinds: unweighted and weighted. Our investigation of these formal models will be motivated by by the Question: What is nondeterminism?

### Leila Amgoud: *Argument-based paraconsistent logics*

Handling inconsistency in propositional knowledge bases (KBs) has been studied in AI for a long time. Several two-level logics have been defined: They start with classical propositional logic and define on top of it a non-monotonic logic that infers non-trivial conclusions from an inconsistent KB. There are at least two families of such logics: coherence-based and argument-based logics. The former compute the set of all maximal (for set inclusion) consistent subbases (MCSs) of a KB, and then they apply an inference mechanism for drawing consequences from the MCSs. Argument-based logics follow another process. They justify every candidate consequence of a KB by arguments, generated using the classical consequence relation, then they identify possible conflicts between arguments, evaluate arguments using a formal method, called semantics, and finally keep among the candidate consequences those that are supported by “strong” arguments.

In this talk, I present three families of argument-based logics that use respectively extension semantics, ranking semantics, and gradual semantics in the evaluation step. I discuss the properties of those logics, and compare them with coherence-based logics.

### Egon Börger: *The Abstract State Machines Method for High Level System Design and Analysis*

We explain the three basic concepts of the Abstract State Machines (ASM) Method for a rigorous development of software intensive systems. The method allows the practitioner to start with an accurate and trustworthy application-domain-centric system model and to link such a `ground model' in a well documented and controllable way through intermediate design steps (called `refinements') to its implementation. The method has been used successfully, under industrial constraints, for the design and analysis of complex hardware/software systems. We highlight some characteristic examples and provide the simple definition of ASMs, starting from scratch. Through its versatility the ASM approach is non-monolithic and integratable at any development level into current software design and analysis environments.