Expressing Ruleset differences as Meta-information for RIF

Hi Allen and RIF members,

          In going through the Design Goals document, I believe that 
an important observation regarding capture of meta-information on 
rule-sets, (which we had addressed as "Supporting the Reuse of Rules" 
use case) has not been included. The last para in section 1 of the 
Design Goals document states the following:


"Besides representing an order of magnitude less effort for the 
implementation of translation algorithms, the use of a RIF has the 
potential for additional benefits. For example, in cases where 
rule-interchange mappings cannot be reliably automated, the 
information provided by the RIF representation should, at the very 
least, be useful in automatically constructing explanations of the 
translation impasse.

It is important to understand, however, that the RIF itself provides 
neither a translation algorithm nor an explicit mapping between rule 
languages. Rather the RIF includes a framework of concepts, 
represented as tags in a markup language, that can be used to provide 
information about the meaning of wffs in a rule language. For rule 
authors who wish to make their rules accessible across languages and 
platforms, the more completely, precisely, and accurately they tag 
their creations using the RIF, the more likely it is that their rules 
will be capable of being automatically translated correctly."

Our use case focuses on the issue of utilizing RIF tags to provide 
meta- information about sets of rules rather than just single rules. 
In my experience while analyzing rules, we have repeatedly 
encountered common patterns across rule sets in various forms, 
regardless of representation and domain that the rules are embedded 
in. In the use case cited in the RIF Use Cases and Requirements, we 
have shown how that different classes such as, MaritimeEquipmentType, 
ElectronicEquipmentType and MiscellaneousEquipmentType have been 
similarly defined in OWL. The last block, in the use case,  abstracts 
the similarities while providing <slots> which indicate the parts 
which are different across the rules. I believe that this type of 
meta-information about sets of rules, such as similarity across the 
rules, can be usefully exploited by systems that utilize RIF for both 
exchange and interoperability of rules.

Below I provide a similar (but shorter) example from Cyc rules from 
the spatial microtheory, that I had analyzed for the DARPA RKF project:

(#$implies
(#$and
(#$termOfUnit ?CONVEXHULLFN (#$ConvexHullFn ?OBJECT))
(#$termOfUnit ?CONVEXHULLFN-1 (#$ConvexHullFn ?CONVEXHULLFN)))
(#$equals ?CONVEXHULLFN ?CONVEXHULLFN-1))

(#$implies
(#$and
(#$termOfUnit ?INTERIORFN (#$InteriorFn ?OBJECT))
(#$termOfUnit ?INTERIORFN-1 (#$InteriorFn ?INTERIORFN)))
(#$equals ?INTERIORFN ?INTERIORFN-1))


This is a case of discovering the characteristic of idempotency 
across various rules where the functions return the same value 
regardless of how many times you invoke it. (termOfUnit in Cyc is 
like a macropredicate to store the value returned by a function call.)

It is important to capture such similarity in invocations, at a 
meta-level in the RIF language, so that these types of rules can be 
invoked in any other language easily and translation process can 
proceed en masse. In other words, we need a way to separate out the 
commonality in formulation of the rules, from the specific 
invocations of these rules, so that these rules can be made more 
amenable to transfer across different rule representations.

So far my attempt to abstracting the commonality has been in the form 
of templates, by recognizing the idempotent rules found above, as a 
UniqueFn, which can invoke the set of rules above by being defined in 
the following manner:

(#$implies
(#$and
   (#$<UniqueFn> ?<UNIQUEFN>)
   (#$termOfUnit ?<UNIQUEFN-TERM-1> (?<UNIQUEFN> ?OBJECT))
   (#$termOfUnit ?<UNIQUEFN-TERM-2> (?<UNIQUEFN> ?<UNIQUEFN-TERM-1>)))
   (#$equals ?<UNIQUEFN-TERM-1> ?<UNIQUEFN-TERM-2>))

Now the exchange across other rule representation languages can be 
carried out both at the syntactic level and at a meta-level  and any 
number of such rules can be invoked, with different names, in either 
one of the languages to be interoperated with.

Thus RIF needs to capture the meta-level formulation of these rules - 
and I see a strong role for advocating first order representation for 
such cases, so that one can invoke rules on sets of classes.

Please note that I am not advocating a first order reasoner (as that 
can be a hard problem to tackle in Phase I) - as I am not executing 
the meta-rules. However it is important to represent them in RIF so 
as to generate other rules.

I would be happy to discuss this further in tomorrow's telecon and/or 
at the F2F in Cannes,

                 Thanks for listening!
                         Mala



At 01:20 PM 2/16/2006, Ginsberg, Allen wrote:

>
>Dear RIF-WGers,
>
>At the last telecon I was assigned the action item of starting to
>compile a list of Design Goal issues on the WIKI.
>
>To view what I have done so far please visit
>http://www.w3.org/2005/rules/wg/wiki/UCR/Design_Goals.
>
>Here is a synopsis:
>
>I list 3 design goal categories.  Each category has some explanation of
>what it is, and one or more of the following
>
>         1) a list of possible design goals
>        2) a list of NOT-A-Design-Goals
>         3) commentary
>         4) questions
>
>My next step is to go back over the messages in the email list and
>build up these lists and add new categories as required. I hope to get
>that done before the next telecon.
>
>Any feedback and suggestions are most welcome.
>
>Allen
>
>
>
>
>
>
>_______________________________________________________________
>
>Dr. Allen Ginsberg        The MITRE Corporation, Information Semantics
>aginsberg@mitre.org       Center for Innovative Computing & Informatics
>
>Voice: 703-983-1604       7515 Colshire Drive, M/S H305
>Fax:   703-983-1379       McLean, VA 22102-7508, USA
>

Mala Mehrotra
Pragati Synergetic Research Inc.  MS 19-46Q, NASA Research Park, 
Moffett Field, CA 94035
Voice:
(650)-625-0274(Office)
(408)-861-0939 (Home Office)
(408)-910-4115 (Cell)
Fax: (408)-516-9599
URL: http://www.pragati-inc.com
Email: mm@pragati-inc.com