Marco Ragni

The challenge of complexity for cognitive systems

Complex cognition addresses research on (a) high-level cognitive processes - mainly problem solving, reasoning, and decision making - and their interaction with more basic processes such as perception, learning, motivation and emotion and (b) cognitive processes which take place in a complex, typically dynamic, environment. Our focus is on AI systems and cognitive models dealing with complexity and on psychological findings which can inspire or challenge cognitive systems research. In this overview we first motivate why we have to go beyond models for rather simple cognitive processes and reductionist experiments. Afterwards, we give a characterization of complexity from our perspective. We introduce the triad of cognitive science methods - analytical, empirical, and engineering methods - which in our opinion have all to be utilized to tackle complex cognition. Afterwards we highlight three aspects of complex cognition - complex problem solving, dynamic decision making, and learning of concepts, skills and strategies. We conclude with some reflections about and challenges for future research.

Cross-Cultural Preferences in Spatial Reasoning

How do people reason about spatial relations? Do people with different cultural backgrounds differ in how they reason about space? The aim of our cross-cultural study on spatial reasoning is to strengthen this link between spatial cognition and culture. We conducted two reasoning experiments, one in Germany and one in Mongolia. Topological relations, such as “A overlaps B” or “B lies within C”, were presented to the participants as premises and they had to find a conclusion that was consistent with the premises (“What is the relation between A and C?”). The problem description allowed multiple possible “conclusions”. Our results, however, indicate that the participants had strong preferences: They consistently preferred one of the possible conclusions and neglected other conclusions, although they were also logically consistent with the premises. The preferred and neglected conclusions were similar in Germany and Mongolia. We argue that the preferences are caused by universal cognitive principles that work the same way in the western culture and Mongolia.

Preferred mental models: how and why they are so important in human reasoning with spatial relations

According to the mental models theory, humans reason by constructing, inspecting, and validating mental models of the state of affairs described in the premises. We present a formal framework describing all three phases and testing new predictions about the construction principle humans normally use and about the deduction process itself - the model variation phase. Finally, empirical findings in support of these principles are reported.

Predicting numbers: an AI approach to solving number series

Solving number series poses a challenging problem for humans and Artificial Intelligence Systems. The task is to correctly predict the next number in a given series, in accordance with a pattern inherent to that series. We propose a novel method based on Artificial Neural Networks with a dynamic learning approach to solve number series problems. Our method is evaluated on an own experiment and over 50.000 number series from the Online Encyclopedia of Integer Sequences (OEIS) database.

Temporalizing spatial calculi: on generalized neighborhood graphs

To reason about geographical objects, it is not only necessary to have more or less complete information about where these objects are located in space, but also how they can change their position, shape, and size over time. In this paper we investigate how calculi discussed in the field of qualitative spatial reasoning (QSR) can be temporalized in order to gain reasoning formalisms that can be used to express spatial configurations and their dynamics. In a first step, we briefly discuss temporalized spatial constraint languages. In particular, we investigate how the notion of continuous change can be expressed in such languages and how continuous change is represented in the so-called conceptual neighborhood graph of the spatial calculus at hand. In a second step, we focus on a special reasoning problem, which occurs quite naturally in the context of temporalized spatial calculi: Given an initial spatial scenario of some physical objects, which scenarios are accessible if the set of all possible paths of these objects is constrained by some further conditions? We show that for many spatial calculi this general problem cannot be dealt with by using the information encoded in the classical neighborhood graphs, as usually discussed in the literature. Rather, we introduce a generalized concept of neighborhood graph, which allows for reasoning about objects in such dynamic settings.

An Arrangement Calculus, Its Complexity and Algorithmic Properties

We define a calculus for spatial reasoning on a grid structure, present a logical calculus, investigate the complexity of the satisfiability problem, we prove its NP completeness and specify additionally a concrete algorithm for solving it.

Dependency calculus: reasoning in a general point relation algebra

The point algebra is a fundamental formal calculus for spatial and temporal reasoning. We present a new generalization that meets all requirements to describe dependencies on networks. Applications range from traffic networks to medical diagnostics. We investigate satisfaction problems, tractable subclassses, embeddings into other relation algebras, and the associated interval algebra.

Solving Raven's IQ-tests: an AI and cognitive modeling approach

Human reasoners have an impressive ability to solve analogical reasoning problems and they still outperform computational systems. Analogical reasoning is relevant in dealing with intelligence tests. There are two kinds of approaches: to solve IQ-test problems in a way similar to humans (i.e., a cognitive approach) or to solve these problems optimally (i.e., the AI approach). Most systems can be associated with one of these approaches. Detailed systems solving geometrical intelligence tests, explaining cognitive operations based on working memory and producing precise predictions and results such as error rates and response times have not been developed so far. We present a system implemented in the cognitive architecture ACT-R, able to solve analogously developed problems of Ravens Standard and Advanced Progressive Matrices. The model solves 66 of the 72 tested problems of both tests. The models predicted error rates correlate to human performance with r = .8 for the Advanced Progressive Matrices and r = .7 for all problems together.

Errare humanum est: Erroneous robots in human-robot interaction

Perfect memory, strong reasoning abilities and flawless performance are typical cognitive traits associated with robots. In contrast, forgetting and erroneous reasoning are typical cognitive patterns of humans. This discrepancy may fundamentally affect the way how robots and humans interact and collaborate together and is today still little explored. In this paper, we investigate the effect of differences between erroneous and perfect robots in a competitive scenario in which humans and robots solve reasoning tasks and memorize numbers. Participants are randomly assigned to one of two groups: in the first group they interact with a perfect, flawless robot, while in the second, they interact with a human-like robot with occasional errors and imperfect memorizing abilities. Participants rate attitude, sympathy, and attributes of the robot in a questionnaire and we measure their task performance. The results show that the erroneous robot triggered more positive emotions but lead to a lower human performance than the perfect one. Effects of both conditions on the group of students with and without technical background are reported.

Spatial conditionals and illusory inferences

ABSTRACT Studies of reasoning often concern specialised domains such as conditional inferences or transitive inferences, but descriptions often cut across such domains, for example: If the circle is to the left of the square then the triangle is to the right of the square. The square is to the right of the circle. The triangle is to the right of the square. Could all three of these assertions be true at the same time? We report four experiments testing the mental model theory of such problems, which combine spatial transitivity and conditional relations. It predicts that reasoners should try to find a single mental model in which all the assertion hold: ○ □ ∆ Such problems should be easier than those that call for a model in which both clauses of the conditional are false, as when the conditional above occurs with: The square is to the left of the circle. The triangle is to the left of the square. In this case, most participants had the “illusion” that the set was inconsistent (Experiment 1). Analogous results occurred when participants evaluated whether a diagram, such as the one above, depicted a possible spatial arrangement (Experiment 2), and when they evaluated the consistency of a conditional and a conjunction (Experiment 3), and of sets of assertions that contained two conditionals (Experiment 4). The findings appear to be beyond the explanatory scope of theories of reasoning based on logical rules or on probabilities.