ICCM: Emotion & Cognition

Cognition and emotion must be partnered in any complete model of a humanlike mind. This article proposes an extension to the Common Model of Cognition – a developing consensus concerning what is required in such a mind – for emotion that includes a linked pair of modules for emotion and metacognitive assessment, plus pervasive connections between these two new modules and the Common Model’s existing modules and links.

This is an in-person presentation on July 20, 2024 (14:00 ~ 14:20 CEST).

Cognitive models are used as simulators that derive external behavior from assumed internal states. As a tool for linking external behavior with internal causes, cognitive models can be used to examine human trait inference on others. While fundamental attribution errors are identified in social psychology, the specific factors remain unclear. By employing detailed cognitive models to specify internal states, it is possible to deepen our understanding of human inference on internal processes. In this study, we utilized the ACT-R cognitive architecture to construct such internal states and externalized behaviors. We also focused on 'curiosity' as an individual trait emphasized in real society to evaluate individuals. We developed a visualizer for the behavior of multiple models of curiosity and conducted subjective evaluations with participants recruited from a Japanese crowdsourcing site. As a result, we observed differences in inferred traits among models, although the specific patterns were not consistently aligned with the model assumptions. Additional analysis revealed that participants' inferences were more influenced by observable behavior patterns rather than internal processes, indicating a deficit in human attribution as suggested by the tradition of social psychology.

This is an in-person presentation on July 20, 2024 (14:20 ~ 14:40 CEST).

For a successful and trustful human-robot interaction the challenge is to provide the robot with information to dynamically adapt appropriately to a person and changing situations. Cognitive architectures such as ACT-R provide valuable capabilities to address this challenge. This paper gives an example on how exactly cognitive architectures can be used to provide better human-robot interaction. First, this paper shows how mental representations can be build up to anticipate the partner and the situation in order to collaborate adaptively. Second, it is shown how to integrate a model with a robot in a simple way. And third, an example is shown how emotion recognition can be used as an example of adapting the interaction accordingly by using perceived changes in the real world. As results the paper gives instructions, concepts and usecase examples on the realization of the different aspects. The paper encourages further research on how cognitive architectures can address challenges in human-aware AI.

This is an in-person presentation on July 20, 2024 (15:00 ~ 15:20 CEST).

We present a computational approach to the mechanisms involved in a type of metacognitive monitoring known as detached mindfulness, a particularly effective therapeutic technique within cognitive psychology. We employ a computational model of metacognitive skill training, founded on the Common Model of Cognition, to articulate the mechanisms through which a detached perception of affect reduces emotional reactivity.

This is an in-person presentation on July 20, 2024 (14:40 ~ 15:00 CEST).