The (newly renamed) Association for the Advancement of Artificial Intelligence (AAAI) is running a (series of) Symposia in Stanford in March ’08, with topics covering interesting areas like:
– AI Meets Business Rules and Process Management
– Architectures for Intelligent Theory-Based Agents
– Semantic Scientific Knowledge Integration
– Symbiotic Relationships between Semantic Web and Knowledge Engineering.
Complex Event Processing technology overlaps with some of these AI topics [*1] in a big way: for example, rule-based CEP technologies expand on the production rule engines developed from expert systems tools for use as business rule engines; underlying distributed event sources and data storage provide efficient blackboard mechanisms for intelligent agents; semantics overlap with the underlying information models of events (especially the correlation of complex events into more meaningful information); etc etc [*2].
As Process Management and Rules are “reasonably” well known to us at TIBCO ( 🙂 ) it might be of interest to study the AI / Business Rules / Process Management symposium (Call for Participation or CfP) in more detail. Disappontingly we don’t get far before detecting hype (i.e. paragraph one), namely that W3C “standards” include RuleML (which is not even part of W3C), and SWRL (which was a “submission”, not a ratified standard). Hmmmm… so why no reference to the “official” W3C standard-in-progress in this area, the W3C Rule Interchange Format working group? Maybe Rule Interchange is considered truly AI-free (although I’d guess the academic and R&D specialists helping craft RIF might disagree).
The remainder of the CfP expands on the semantics theme – the relationship between nascent OMG SBVR and AI semantics research, for example. This is interesting, but disappointingly limits today’s AI research to W3C’s vision of a semantic web (although one could argue that either has hijacked the other). Relevant AI topics that didn’t get a mention include:
– machine definition of processes (such as automated generation of BPMN diagrams of processes)
– machine generation (induction) of policy-type business rules (as in SBVR business constraints) – note that rule induction of classification rules for production rule engines is also a long running (and worthy) AI topic, which would be even more useful given standards like (also not mentioned) OMG PRR
– AI techniques for Complex Event Processing [ah, you knew we’d get back on topic at some point…].
Maybe next time!
Notes:
[1] One paper directly of interest to the CEP community is “Rules for Making Sense of Events: Design Issues for High-Level Event Query and Reasoning Languages” by Francois Bry and Michael Eckert.
[2] More examples on AI overlapping with CEP: this comment on distributed agents, or compare this commentary in the Wiki entry on the history of AI with these comments on what is needed in CEP…
