U.S. Department of Energy

Pacific Northwest National Laboratory

SHyRe Highlighed at Smart Data Conference

Abstract

We will introduce three national lab use cases that require efficient reasoning on large and diverse data streams. Attendees will see examples of challenges including sampling rates and cache maintenance, deductive reasoning with partial information. We will introduce a semantic web framework that addresses challenges around (a) choosing assertions worthy of preserving within a fixed-size cache and (b) performing deductive reasoning continuously using a “smart cache” in order to effectively draw relevant conclusions as quickly as possible. The use of a smart cache differentiates our work from state-of-the-art works in deductive stream reasoning in that the cache enables us to temporarily store propositions that are no longer in the stream window. We will present a detailed first hand case study of our prototypes that use 3 different semantic web reasoners and describe the benefits in concrete application terms, thereby providing testimonials for the approach.

Smith WP, and DL McGuinness.  2015.  "The PNNL Streaming HYpothesis REasoning (SHYRE) Use Case."  Abstract submitted to Smart Data Conference, San Jose, CA.  PNNL-SA-109299. 

| Pacific Northwest National Laboratory