Second Workshop on Human Reasoning and Computational Logic
The slides will be online. You can find them under Talks.
From the 10th to the 12th of April 2018, we organize the second workshop on Human Reasoning and Computational Logic at TU Dresden, Germany. The goal of this workshop is to provide a platform for the scientific exchange with respect to Human Reasoning between the areas of Cognitive Science and Computational Logic. For the first workshop on Human Reasoning and Computational Logic at TU Dresden, see here.
Dates
- 10th to 12th of April, 2018
- Participation fee is 50EUR, please contact Emmanuelle Dietz for details
- Dinner at Maiwerts, a group will meet at 18:00 in front of the faculty to leave together taking the tram
Venue
The workshop is held at the Computer Science Faculty building of Technische Universität Dresden, Nöthnitzer Straße 46, Dresden-Räcknitz .
How to reach us
Contact person
Reasoning and Consciousness---Teaching a Theorem Prover to let its Mind Wander by Claudia Schon and Ulrich Furbach (Universität Koblenz-Landau) slides
Iterated contraction of propositions and conditionals under the principle of conditional preservation by Gabriele Kern-Isberner (Technische Universität Dortmund), Tanja Bock (Technische Universität Dortmund), Kai Sauerwald (FernUniversität in Hagen) and Christoph Beierle (FernUniversität in Hagen) slides
Research on iterated belief change has focused mostly on belief revision, only few papers have addressed iterated belief contraction. Most prominently, Darwiche and Pearl published seminal work on iterated belief revision the leading paradigm of which is the so-called principle of conditional preservation. In this paper, we use this principle in a thoroughly axiomatized form to develop iterated belief contraction operators for Spohn’s ranking functions. We show that it allows for setting up constructive approaches to tackling the problem of how to contract a ranking function by a proposition or a conditional, respectively, and that semantic principles can also be derived from it for the purely qualitative case.
Using State-Transition Systems for Reasoning about Cognitive Architectures by Kai Sauerwald (FernUniversität in Hagen), Marco Ragni (Albert-Ludwigs-Universität Freiburg), Tanja Bock (Technische Universität Dortmund), Gabriele Kern-Isberner (Technische Universität Dortmund), and Christoph Beierle (FernUniversität in Hagen) slides
Cognitive architectures, like ACT-R, SOAR or Clarion, are software systems that are used as a platform for cognitive modelling. Besides of their usage as a tool for modelling, each cognitive architecture represents a cumulation of work on the invariant structures of the mind. This makes a formal analysis of their properties attractive. For instance, cognitive architectures could contain inherent knowledge about the constraints of human reasoning. Unfortunately the analysis is difficult, because most of the architectures are defined in a rather informal way. In this talk, we present a formal framework based on transition systems providing a language to describe the computational processes of cognitive architectures by abstracting from respective implementation details. This provides a formal common ground for reasoning about the computational processes that are realized in different cognitive architectures. Notions of equivalences enable the comparison of cognitive models and architectures on diverse levels of detail.
Syllogistic Reasoning in Seven Spaces by Frieder Stolzenburg, joint work with Raimund Lüderitz (Hochschule Harz) slides
Syllogisms and syllogistic reasoning has been the subject of research and scientific discourse for more than two millennia. Syllogisms sum quantified assertions into an overall statement, usually consisting of two premises and one conclusion. While syllogistic reasoning can be modeled by classical first-order logic in a straightforward manner, it is an open question which of the possible syllogisms are accepted as valid by human reasoners. In this talk, we present an approach that models the reasoning process with seven spaces of a set diagram. It can easily be implemented by constraint logic programming. We distinguish several assumptions that humans may make during their reasoning process, in particular that all used categories are non-empty. In contrast to pure logic-based approaches, the proposed procedure allows to represent diverse models human reasoners may follow. The results show good correlation and coincidence with psychological investigations.
The WCS Approach in the Syllogism Challenge: Reasoning Principles and their Cognitive Adequacy by Emmanuelle-Anna Dietz (Technische Universität Dresden) slides
A computational logic approach based on the Weak Completion Semantics participated at the Syllogism Challenge 2017. This approach provides a formalization of various cognitive principles and characterizes different human reasoning clusters, i.e. groups of reasoners that diverge in their conclusions. In this talk, a brief introduction of this approach is provided. After that, the adequacy of the formalization of two cognitive principles, contraposition and unknown generalization, is discussed.
Cognitive Principles and Reasoning Clusters in Multinomial Process Tree: A Case Study for Human Syllogistic Reasoning by Orianne Bargain (Technische Universität Dresden) slides
A syllogism is a kind of logical argument that applies deductive reasoning to arrive at a conclusion based on two or more propositions that are asserted or assumed to be true. It was first defined by Aristotle and then investigated by psychologists over 100 years ago. Various psychological experiments have shown that participants did not always derive the classical logical valid conclusions given two syllogistic premises. Recently a new cognitive theory based on the Weak Completion Semantics has been developed and seems to give good results. A recent extension of this approach tries to identify clusters of reasoners instead of modeling the human reasoner. In this presentation, we investigate the representation of these reasoning clusters and cognitive principles by means of multinomial process tree models.
Programming Machine Ethics under the Weak Completion Semantics by Dominic Deckert and Maximilian Schlosser (Technische Universität Dresden)
While computer-driven systems become more ubiquitous and complex, truly autonomous systems and robots will surely emerge in the near future. But this development is not without concerns, one of which is the amount of trust we can put in such systems. This is where Machine Ethics comes into play: finding ways to ensure the morality of autonomous behaviour, and how different approaches to ethical behaviour can be provided to such a system. The Trolley Problem, an important test case for Machine Ethics, is the focus of this presentation. We discuss some moral reasoners for variants of the Trolley Problem under the Weak Completion Semantics (WCS). Furthermore, we present a unified formalism for describing Trolley Problem variants and a translation algorithm into a WCS-based logic program.
Obligation and Factual Conditionals: The Suppression Task Revisited by Steffen Hölldobler (Technische Universität Dresden) slides
Ruth Byrne has introduced obligation and factual conditionals in [1]. In [2] obligation and factual conditionals were modelled under the weak completion semantics and several human reasoning tasks were solved adequately under this semantics. In this talk I will revisit the suppression task [3] and investigate whether the weak completion semantics models this task adequately if the conditionals are classified as obligation or factual conditionals.
[1] R. M. J. Byrne: The Rational Imagination: How People Create Alternatives to Reality. MIT Press: 2007
[2] E.-A. Dietz Saldanha, S. Hölldobler I. Louredo Rocha: Obligation versus Factual Conditionals under the Weak Completion Semantics. In Proc. YSIP2, http://ceur-ws.org/Vol-1837/, 55-64, 2017
[3] R. M. J. Byrne: Suppressing Valid Inferences with Conditionals. Cognition 31, 61-83: 1989
Inhomogeneity in Reasoning: A Challenge for Cognitive Modeling by Lukas Elflein (Albert-Ludwigs-Universität Freiburg) slides
Human syllogistic and conditional reasoning differs systematically from classical logical frameworks. Multiple computational theories describe these deviations based on aggregates of experimental data. We show that their description of inter-individual differences in reasoning is very limited.
Also, we show that they underperform in predicting individual human behavior. We discuss implications for cognitive modeling and future developments.Each talk should be not more than 40 minutes long, with at least 5 minutes of the slot reserved for Q&A.
Tuesday, 10.4.2018
13:30 - 14:15 Obligation and Factual Conditionals: The Suppression Task Revisited by Steffen Hölldobler
14:15 - 14:45 Coffee break
14:45 - 15:30 Reasoning and Consciousness---Teaching a Theorem Prover to let its Mind Wander by Claudia Schon and Ulrich Furbach
15:30 - 16:00 Coffee break
16:00 - 16:45 Inhomogeneity in Reasoning: A Challenge for Cognitive Modeling by Lukas Elflein
Wednesday, 11.4.2018
9:30 - 10:15 Iterated Contraction of Propositions and Conditionals under the Principle of Conditional Preservation by Gabriele Kern-Isberner, Tanja Bock, Kai Sauerwald, and Christoph Beierle
10:15 - 10:45 Coffee break
10:45 - 11:30 Using State-Transition Systems for Reasoning about Cognitive Architectures by Kai Sauerwald, Tanja Bock, Gabriele Kern-Isberner and Christoph Beierle
11:30 - 13:30 Lunch break
13:30 - 14:15 Syllogistic Reasoning in Seven Spaces by Frieder Stolzenburg
14:15- 14:45 Coffee break
14:45 - 15:30 The WCS Approach in the Syllogism Challenge: Reasoning Principles and their Cognitive Adequacy by Emmanuelle-Anna Dietz
15:30 - 16:00 Coffee break
16:00 - 16:30 Cognitive Principles and Reasoning Clusters in Multinomial Process Tree: A Case Study for Human Syllogistic Reasoning by Orianne Bargain
19:30 Dinner
Thursday, 12.4.2018
9:30 - 10:15 Programming Machine Ethics under the Weak Completion Semantics by Dominic Deckert and Maximilian Schlosser
10:15 - 10:45 Coffee break
10:45 - 11:30 tba
