Abstract: This talk will address computational cognitive vision at the interface of (spatial) language, (spatial) logic, (spatial) cognition, and artificial intelligence. I will present general methods for the semantic interpretation of dynamic visual imagery with an emphasis on the ability to perform grounded reasoning and learning with cognitively rooted structured characterizations of commonsense knowledge pertaining to space, action, and change. Focus will be on deep semantics encompassing: (1) General foundational commonsense abstractions –e.g., of space and motion– necessary for `grounding’ dynamic visual stimuli; (2) Systematically formalised declarative reasoning and learning with commonsense abstractions pertaining to space, space-time, motion, actions & events, and spatio-linguistic conceptual knowledge; and (3) Human-centred interpretable models of computational visuospatial commonsense integrating symbolic and neural techniques geared towards high-level capabilities such as space-time mental model formation, counterfactual reasoning, perceptual narrativisation. Methods developed and overall approach positioned are intended to serve as a systematic model and general framework for integrating diverse, multi-faceted AI techniques pertaining knowledge representation and reasoning, computer vision, and machine learning towards realising a human-centric model of computational visual intelligence supporting Embodied AI encompassing (grounded) reasoning, learning, and planning. In this respect, I will highlight (one or more) applications of cognitive vision in autonomous driving, cognitive robotics, cognitive media studies (relevant to visual arts and media studies), and visual perception / cognitive psychology. More broadly, I hope to emphasise an interdisciplinary approach –at the confluence of Artificial and Human Intelligence– that is needed to better appreciate the complexity and spectrum of varied human-centred challenges in the design and (usable) implementation of machine (visual) intelligence in diverse human-system interaction contexts likely to also require normative compliance vis-à-vis emerging AI regulation.

BIOGRAPHY: Mehul Bhatt is Professor within the School of Science and Technology at Örebro University (Sweden). His research addresses the confluence of ‘Artificial and Human Intelligence’, investigating the formal, cognitive, and computational foundations of next-generation human-centred AI and cognitive technologies with a principal emphasis on human-centred knowledge representation, semantics, computational commonsense reasoning, integration of (machine) reasoning & learning, and declarative spatial reasoning. A key recent focus has been on (neurosymbolic) reasoning & learning about space, action and motion in human-scale embodied multimodal interaction. Mehul Bhatt steers CoDesign Lab (www.codesign-lab.org), an initiative aimed at addressing the confluence of Cognition, Artificial Intelligence, Interaction, and Design Science for the development next-generation Human-Centred AI and Cognitive Technologies, and the human-centred design, development, and usability of cognitive technologies aimed at ‘human-in-the-loop’ assistance & empowerment in decision-making, planning, creative-analytical problem-solving, and automation. Since 2014, he also directs the research and consulting group DesignSpace (www.design-space.org), and pursues ongoing research in Cognitive Vision and Perception (www.codesign-lab.org/cognitive-vision).

Mehul Bhatt obtained a bachelors in economics (India), masters in information technology (Australia), and a PhD in computer science (Australia). He has been a recipient of an Alexander von Humboldt Fellowship, a German Academic Exchange Service award (DAAD), and an Australian Post-graduate Award (APA). He was the University of Bremen nominee for the German Research Foundation (DFG) Award: Heinz Maier-Leibnitz-Preis 2014. Previously, Mehul Bhatt was Professor at the University of Bremen (Germany). Details are available via: https://mehulbhatt.org.