Uploaded October 2024 | Updated September 2026, 20 hours ago
Recorded as part of the CFvW Colloquium on October 16, 2024
Multimodality in natural and artificial intelligence: Philosophical Perspectives - PD Dr. Jens Lemanski (Universität Münster & FernUniversität in Hagen
Talk abstract:
The concept of multimodality is currently being explored in numerous fields of intelligence research, including animal and human communication, as well as machine learning. One of the most significant challenges currently being addressed in all areas of intelligence research is that of reasoning. The question thus arises as to whether there exists a modality type, such as text-based approaches, that is afforded a privileged access in the domain of reasoning. This question is linked to the assumption that visual and tactile modalities, such as diagrams or gestures, can process less complex information more efficiently and have a lower degree of conventionalisation than other types of modality. However, empirical evidence indicates that reasoning tasks are solved faster and more effectively not only by natural beings but also by machine agents when multimodal rather than unimodal solution strategies are employed.
The lecture addresses two key aspects: firstly, it considers the philosophical approach that determines the choice between unimodal and multimodal reasoning; secondly, it demonstrates that there have been numerous approaches in the history of philosophy that have advocated multimodal reasoning and have drawn different boundaries of complexity and conventionalisation for visual and tactile modalities. These insights have the potential to inform and enhance contemporary research into intelligence.
Explore our colloquium schedule on our website: uni-tuebingen.de/en/research/centers-and-institutes/carl-friedrich-von-weizsaecker-center/news-and-events/carl-friedrich-von-weizsaecker-colloquium
Recorded as part of the CFvW Colloquium on October 16, 2024
Multimodality in natural and artificial intelligence: Philosophical Perspectives - PD Dr. Jens Lemanski (Universität Münster & FernUniversität in Hagen
Talk abstract:
The concept of multimodality is currently being explored in numerous fields of intelligence research, including animal and human communication, as well as machine learning. One of the most significant challenges currently being addressed in all areas of intelligence research is that of reasoning. The question thus arises as to whether there exists a modality type, such as text-based approaches, that is afforded a privileged access in the domain of reasoning. This question is linked to the assumption that visual and tactile modalities, such as diagrams or gestures, can process less complex information more efficiently and have a lower degree of conventionalisation than other types of modality. However, empirical evidence indicates that reasoning tasks are solved faster and more effectively not only by natural beings but also by machine agents when multimodal rather than unimodal solution strategies are employed.
The lecture addresses two key aspects: firstly, it considers the philosophical approach that determines the choice between unimodal and multimodal reasoning; secondly, it demonstrates that there have been numerous approaches in the history of philosophy that have advocated multimodal reasoning and have drawn different boundaries of complexity and conventionalisation for visual and tactile modalities. These insights have the potential to inform and enhance contemporary research into intelligence.
Explore our colloquium schedule on our website: uni-tuebingen.de/en/research/centers-and-institutes/carl-friedrich-von-weizsaecker-center/news-and-events/carl-friedrich-von-weizsaecker-colloquium
