Magazine article AI Magazine

Qualitative Reasoning: Everyday, Pervasive, and Moving Forward-A Report on QR-15

Magazine article AI Magazine

Qualitative Reasoning: Everyday, Pervasive, and Moving Forward-A Report on QR-15

Article excerpt

The 28th International Workshop on Qualitative Reasoning (QR-15) presented advances toward reasoning tractably with massive qualitative and quantitative models, automatically learning and reasoning about continuous processes, and representing knowledge about space, causation, and uncertainty.

**********

The 28th International Workshop on Qualitative Reasoning (QR-15) was held in Minneapolis, MN on August 10-11, 2015. The technical track included two invited talks, 11 oral presentations, and 5 poster presentations. All of the research presented at QR-15 utilized qualitative representations of space, physical systems, causality, or uncertainty, but the domains of application were diverse: we saw a qualitative model of uncertainty demonstrated in a cybersecurity domain, which was robust to sensor noise and incomplete knowledge (Robert Goldman, SIFT); we learned about extensions to frame semantics for modeling continuous processes (Clifton McFate, Northwestern); we discussed multiple strategies for encoding topological relations and generating proofs over qualitative spatial descriptions; we learned about novel applications of process-oriented modeling to ecology and limnology (Peter Struss, TUM); we heard news of terabyte-scale qualitative simulations to automatically vet designs of cyber-physical systems (Johan de Kleer, PARC); and we discussed qualitative graph abstraction techniques (Scott Friedman, SIFT) that may ultimately reduce the size of these qualitative simulations and thereby improve efficiency.

Kenneth Forbus (Northwestern) delivered an invited talk on qualitative reasoning in the everyday world. Forbus's talk focused on the pervasiveness of qualitative knowledge and qualitative reasoning in everyday cognition. Elementary-level science exams --some of which are used to assess the reasoning capabilities of AI systems--are composed of problems that demand qualitative reasoning and knowledge of continuous processes and qualitative proportionality. This is evidence that qualitative reasoning is a central component of everyday human cognition and formal education, and it suggests that qualitative reasoning should be a key focus for building human-level artificial intelligence. This pervasiveness of qualitative reasoning motivates the community's work on automatically building or acquiring qualitative models of the world, such as by experiential learning or reading science texts. Clifton McFate (Northwestern) presented work on frame semantics of continuous processes, which supports these goals of learning qualitative models from texts.

Johan de Kleer (PARC) delivered an invited talk on lessons learned from the Adaptive Vehicle Make program. The talk focused on the challenges of model-based reasoning and qualitative simulation at massive scale. One such challenge is prioritizing the expansion of the state space in a purely structural qualitative or quantitative model. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.