Research projects

Artificial intelligence (AI) has already made possible the communication between machines and humans, in addition to providing solutions to problems. But could AI grasp deeper meanings or motivations? The EU-funded MUHAI project will investigate by exploring a radically new approach in studies on human-centric AI technology focussing on meaning and understanding. Researchers will refer to the theory of personal dynamic memory to shed light on the process of reconstructing meanings of experiences. Based on recent advances in technical AI components, such as deep learning networks, it will build new ones, such as the mental simulation of actions using gamified virtual environments. The new approach will be tested in cases that require common sense and an understanding of social phenomena, as for example the persistence of inequality in our society.

EASE is an interdisciplinary research center at the University of Bremen that investigates everyday activity science and engineering. Its core purpose is to advance our understanding of how human-scale manipulation tasks can be mastered by robotic agents. To achieve this, EASE establishes the research area “Everyday Activity Science and Engineering” and creates a research community that conducts open research, open training, open data, and knowledge sharing

This study group engages with the challenge of modeling conceptual knowledge and conceptual change, in the specific domain of science. Our engagement with this challenge is motivated by two key practical goals: helping learners develop understanding of scientific concepts in the context of formal education; and promoting the public understanding of science.

There are many strands of research across a number of related disciplines that have made contributions to our understanding of the nature of conceptual knowledge and conceptual change in science. These research strands span cognitive and developmental psychology, history and philosophy of science, computer science, science studies, the learning sciences, educational psychology and others. As empirical and theoretical research accumulates across these multiple bodies of work some foundational constructs are understood in different ways and core debates are engaged in without sufficient precision.

Through the work of this study group we hope to clarify the following foundational constructs and the relationships between them: concepts, beliefs, image schemas, mental models and skeletal causal frameworks. We also hope to fine-tune debate over whether naïve knowledge is best characterized as coherent or fragmented, whether we must sometimes characterize conceptual change as involving transformations between incommensurable conceptual structures, and over the roles of epistemic beliefs about knowledge and trust in other people in belief acceptance and the processes inducing conceptual change.

We start our work with the assumption that to achieve these goals three perspectives must be brought into dialogue: cognitive perspectives on mental structures and processes, computational perspectives on cognitive modeling, and sociohistorical perspectives on knowledge and knowledge change. The three conveners of this study group represent these three perspectives. The series of meetings will be designed to explore how these three perspectives combine productively and constrain one another in the effort to clarify the foundational constructs and fine tune the core debates mentioned above. As an outcome, the study group aims to provide a more solid theoretical grounding and motivate the needed empirical research to improve our understanding of conceptual knowledge and conceptual change in the context of formal science instruction and the public understanding of science.

Information is the foundation of all rational decision-making. Without suitable Information Systems, individuals, organizations, communities and governments can neither systematically take optimal decisions nor understand the full effect of their actions. information system (IS) contain information structures that are expected to represent portions of reality. For this reason, the quality of an IS strongly and directly depends on the faithfulness of these representations w.r.t. the aspects of reality they are supposed to represent. Increasingly, ISs are being developed in complex critical domains where a mistaken interpretation of the meaning of a piece of information may have serious social consequences.

In such domains, ISs engineering cannot succeed without a proper ontology-driven conceptual modelling (ODCM) support for producing models of information structures that are both ontologically accurate (i.e., expressive and precise enough to characterize important distinctions of the worldview being described), and ontologically consistent (i.e., able to capture an aggregation of worldviews that are consistent with each other). Over the years, there has been a remarkable scientific progress on ODCM as well as an increasing recognition of its relevance to IS engineering practice. However, most efforts in this area are limited to the ontological category of endurants (objects, things), neglecting the complementary category of Events (perdurants, happenings, processes, occurrences). The overall goal of this project is to develop theoretical results and engineering tools to support the analysis of event-centric phenomena in critical domains and to enable the representation of events in structural (entity-centric) conceptual models. The results of this project will be evaluated in the domain of public safety (violent crimes).

Embodied cognitive science has revolutionised the way we today understand the workings of the human mind. The cognitive turn in Artificial Intelligence (AI) is also in full swing, however significant challenges are raised by a lack of understanding of how to analyse formally key concepts of cognition. Also lacking is a sufficient understanding of the ontology of these key concepts, and indeed an agreement on even basic definitions and their formal counterparts.

The core aim of \project is to initiate a systematic study into the formal logical modelling of the building blocks of human conceptual thinking, focusing on the deeply interlinked notions of affordance and image schema. The main challenge here is that the common-sense abstractions of these notions deviate significantly from the typical spatial and temporal models found in contemporary mainstream knowledge representation. To this end, we will study novel conbinations and variations of models of time and space, and devise corresponding cognitively motivated spatio-temporal logics.

The key deliverable of \project, thus, will be to provide a portfolio of spatio-temporal logics suitable to modelling the phenomena encountered in basic cognitive modelling, and to validate these formalisms in appropriate case studies and empirical validation.

The work will be closely co-ordinated with partners working in Cognitive AI and Image Schema modelling, namely Artificial Intelligence Research Institute (Spain), University of Osnabr\”uck (Germany), University of Ni\v{s} (Serbia), University College Dublin (Ireland), University of Magdeburg (Germany), and University of Bremen (Germany).

Scientific exchange program between Free University of Bozen-Bolzano, Italy, and the University of Bremen, Germany.

Internally funded research project.

The LOGIS project, running between 2015-17, focuses on reaching out into a core topic in cognition: the logic of image schemas. The theory of image schemas was developed within cognitive linguistics, and has been extremely influential since Lakoff & Johnson’s work in the late 1980s. Although often acknowledged as central for the realisation of artificial creative systems, the difficult interplay between the deep roots of image schemas in human cognition and the necessity to (logically) formalise them for computational systems, is largely unexplored. This problem is investigated on the theoretical, logical level, and extensive empirical work will be carried out to verify the viability of certain formal distinctions. The LOGIS project is co-ordinated with University College Dublin (Ireland) and the University of Nis (Serbia). Experiments and field trips are carried out in the multi-lingual regions of Alto-Adige as well as Nis.

Internally funded research project on the relationship between applied ontology and concept invention.

COINVENT is a high-profile international research project funded by the European Commission’s 7th Framework Programme, and starting October 2013. It aims at advancing the formal understanding of creativity by developing a computationally feasible, cognitively-inspired formal model of concept invention, drawing from interdisciplinary research results from cognitive science, artificial intelligence, formal methods and computational creativity, and validating it for mathematical reasoning and melodic harmonisation. The project is coordinated by Dr. Marco Schorlemmer at IIIA-CSIC Barcelona.