Doctoral Seminar: Ammar Alsaig

Speaker: Ammar Alsaig

Supervisors: Drs.  V. S. Alagar, N. Shiri

Supervisory Committee: Drs.  J. Bentahar, D. Goswami, G. Grahne

Title: Contelog:  A Framework for Representing and Reasoning With Contexts

Date: Thursday, March 28, 2019

Time: 10:15 a.m.

Place: EV 3.309

ABSTRACT

The notion of context has been around for a long time and has been the subject of numerous research by logicians, philosophers, and linguists. This in turn has resulted in the development of concepts,  techniques, and tools for modeling and reasoning with contexts. Many modern day computing applications, especially in areas such as ubiquitous and personalized services, and autonomous vehicles, have  emphasized the need for a rigorous development of context-aware  systems whose behaviours can be formally analyzed. However, existing  methodologies and techniques are mostly ad-hoc as they implement the  required methods with no formal representation for context, and more  importantly the tendency has been to loosely put together  deep-learning algorithms with knowledge-base system techniques. Such loose assemblage lack a solid formal foundation to support the systematic development, reasoning, and maintenance of context-aware systems. In this talk, after briefly surveying the current state of  research on contexts and contextual reasoning, the new framework  Contelog which is formal, flexible, and scalable is introduced.  Contelog is a logical framework for context modeling and reasoning, in  which contexts are represented and treated as first class citizens.  The proposed framework conservatively extends the syntax of Datalog, originally proposed for deductive database systems, and provides a  declarative fixpoint semantics of programs in which facts and rules  are annotated with contexts. The expressive power of Contelog has been illustrated in a Book of Contelog Examples, from which an interesting example is chosen to demonstrate the reasoning power of Contelog.