Affiliation: University of Edinburgh
Personal Website: https://vaishakbelle.com
Title: Tractable Probabilistic Models for Ethical and Causal AI
Abstract: Among the many ethical dimensions that arise in the use of ML technology, three stand out as immediate and profound: enabling the interpretability of the underlying decision boundary, addressing the potential for learned algorithms to become biased against certain groups, and capturing blame and responsibility for a system’s outcomes. In this talk, we advocate for a research program that seeks to bridge tractable (probabilistic) models for knowledge acquisition with rich models of autonomous agency that draw on philosophical notions of beliefs, intentions, causes and effects.
Affiliation: University of Münster
Personal Website: https://www.uni-muenster.de/Geoinformatics/en/sil/index.html
Title: Qualitative Spatial Evaluation of Sketch Maps
Abstract: Sketch Mapping, i.e. drawing a map on a piece of paper, is a common method to investigate human spatial knowledge. To evaluate, how complete, correct and generalized a person’s spatial knowledge is, we analyze qualitative spatial relations among the sketched features and compare them to relations between corresponding features in city maps. Since the sketched information is often generalized, our research investigates how generalization and incompleteness affect qualitative spatial relations at different levels of abstraction.
John F. Sowa and Arun K. Majumdar
Personal Website: http://www.jfsowa.com
Title: Neuroconceptual Processing for Learning, Reasoning, and Language Understanding
Short explanation: The term neurosymbolic has become popular for methods that combine neural networks with symbolic methods. Our neuroconceptual system combines arbitrarily large graphs of any kind, including conceptual graphs, with a system that can do learning as fast or faster than any neural nets. But more importantly, it supports high-speed logic on the same graphs used for learning and language processing. Since conceptual graphs are the basis for what the system learns, it can explain what it learns by translating the graphs to and from natural languages.