---
title: "Massively Parallel Support for Efficient Knowledge Representation"
authors:
  - "Matthew P. Evett"
  - "William A. Andersen"
  - "James A. Hendler"
doi: ""
url: "https://ijcai.org/proceedings/1993/183"
pdf: "https://ijcai.org/proceedings/1993/0183.pdf"
venue: "Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence"
year: 1993
pages: "1325-1330"
track: "Neural Networks"
keywords:
---

# Massively Parallel Support for Efficient Knowledge Representation

Matthew P. Evett, William A. Andersen, James A. Hendler

*Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence*, pages 1325-1330.

## Abstract

PARKA is a frame-based knowledge representation system implemented on the Connection Machine. The system, which provides extremely fast property inheritance inference capabilities, performs fast "recognition" queries of the form "find all frames satisfying m property constraints " in O(d+m) time-proportional only to the depth of the knowledge base (KB), and independent of its size. PARKA's performance is measured in tenths of a second, even for KBs with 100,000+ frames. We show similar results on the Cyc KB, and for queries involving knowledge structure pattern matching in support of case-based planning. With such run-time performance, PARKA is possibly the "fastest knowledge representation system in the world".