D. A. McAllester. Reasoning utility package user's manual. Memo 667, MIT Artificial Intelligence Lab., April 1982.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....extensions that improve its performance. Section 5 concludes with a discussion of alternate approaches and related work. 2 Background and Environment The completeness problem described briefly above first came to our attention when we began to use McAllester s Reasoning Utility Package (tuP) [5] in our research related to reasoning about programs [8] Among other facilities, RUP included equality reasoning, a primitive pattern directed invocation mechanism (demons were associated with operator symbols) and a truth maintenance system. We used these facilities to build up a library of ....

D. A. McAllester. Reasoning utility package user's manual. Memo 667, MIT Artificial Intelligence Lab., April 1982.

....by transitivity from other elements. The algebra of sets (lines 15 17) involves the theory of membership, subset, union, intersection and complements. The propositional layer and the congruence closure algorithm of the algebraic layer are derived from McAllester s Reasoning Utility Package [3]. Reasoning about equality is invoked several times in the Requirements Apprentice scenario to detect potential problems. Other algebraic properties, such as transitivity, commutativity, etc. come up everywhere in formal modeling tasks. The frames layer, which is built using facilities from ....

D. A. McAllester. Reasoning Utility Package User's Manual. Memo 667, MIT Artificial Intelligence Lab., April 1982.

....extensions that improve its performance. Section 5 concludes with a discussion of alternate approaches and related work. 2 Background and Environment The completeness problem described briefly above first came to our attention when we began to use McAllester s Reasoning Utility Package (RUP) [5] in our research related to reasoning about programs [8] Among other facilities, RUP included equality reasoning, a primitive pattern directed invocation mechanism (demons were associated with operator symbols) and a truth maintenance system. We used these facilities to build up a library of ....

D. A. McAllester. Reasoning utility package user's manual. Memo 667, MIT Artificial Intelligence Lab., April 1982.

....processes such as Horn clause inference and congruence propagation. This is accomplished through a propositional knowledge base that supports a tractable version of forward chaining (Vilain, 1995) along with an equality algorithm based on congruence closure. Alembic inherits this architecture from McAllester s (1982) mythic RUP. The approach allows the system to decouple syntacticallydriven front end processing, i.e. name tagging and parsing, from the back end processing required for extraction, i.e. co reference, event merging, identification of extractable entities or relations, domain independent ....

McAllester, D. (1982). Reasoning Utility Package User's Manual. AI Memo 667, MIT AI Lab.

....extensions that improve its performance. Section 5 concludes with a discussion of alternate approaches and related work. 2 Background and Environment The completeness problem described briefly above first came to our attention when we began to use McAllester s Reasoning Utility Package (rup) [9] in our research related to reasoning about programs [12] Among other facilities, rup included equality reasoning, a primitive pattern directed invocation mechanism (demons were associated with operator symbols) and a truth maintenance system. All of these operated on a data base of ground ....

D. A. McAllester. Reasoning utility package user's manual. Memo 667, MIT Artificial Intelligence Lab., April 1982.

....being employed for large object bases. In the context of terminological systems, an early solution to the retraction problem was provided by KL TWO [Vil85] It consisted of a purely terminological component (not supporting object management) and, as its assertional component, a version of RUP [McA82] a reason maintenance system supporting propositional logic. Although an elegant solution with respect to the retraction problem, the two components did not match very well (cf. Neb90, p. 64] for details) The BACK system is more closely related to the systems CLASSIC of Bell Labs, and LOOM of ....

David A. McAllester. Reasoning utility package user's manual. AI Memo 667, AI Laboratory, Massachusetts Institute of Technology, Cambridge, Mass., April 1982.

....satisfiability (by restricting the propositional language to Horn logic, for instance) then unambiguous base revision will be itself polynomial. This result gives a formal justification for the claim made in (Nebel 1989) that this form of revision is similar to the functionality the rup system (McAllester 1982) offers in an abstract sense, though. 10 The important point to note is that a feasible implementation of belief revision is possible if we restrict ourselves to polynomial methods for satisfiability by restricting the language or by tolerating incompleteness and by using a polynomial method ....

McAllester, D. A. (1982): Reasoning Utility Package User's Manual, AI Memo 667, AI Laboratory, Massachusetts Institute of Technology, Cambridge, Mass.

....member (PROTEUS) 8 NIKL does define individual class frames as classes that have only a single member but classes nonetheless. To reason about individual objects a user must employ an additional system, for example KL TWO [98] combines NIKL with a propositional reasoning system called RUP [54]. ffl Descriptions These frames are variablized classes that are employed in the UNIT Package to represent goals in planning problems [93] CLASS also has indefinites. ffl Prototypes KRL, RLL, and JOSIE employ prototype frames to represent information about a typical instance of a ....

D.A. McAllester. Reasoning utility package user's manual. Technical Report AI Memo 667, Massachusetts Institute of Technology AI Laboratory, 1982.

....For instance, learning [ Diettrich, 1986 ] diagnosis from first principles [ Reiter, 1987 ] and interpretation of counterfactuals [ Ginsberg, 1986 ] are all activities requiring the revision of beliefs. In Artificial Intelligence, a number of so called truthmaintenance systems [ Doyle, 1979; McAllester, 1982; This paper describes research done at the Technical University of Berlin as part of ESPRIT project 311 and at IBM Germany as part of the LILOG project. de Kleer, 1986 ] were developed which support belief revision. However, the question remains how belief revision can be described on an ....

.... reconstruction, contrary to the opinion most authors seem to have (cf. Ginsberg, 1986; Gardenfors, 1988 ] Finally, in Section 7, we will refine the reconstruction in order to satisfy all rationality postulates leading to a belief revision strategy similar to the one used in the rup system [ McAllester, 1982 ] 2 The Theory of Epistemic Change For the following discussion, we will assume a propositional language L containing propositions x; y; z and the standard sentential connectives ( Sets of propositions will be denoted by A; B; C. Furthermore, shall denote classical ....

[Article contains additional citation context not shown here]

David A. McAllester. Reasoning Utility Package User's Manual. AI Memo 667, Cambridge, Mass.: AI Laboratory, MIT, 1982.

....orderings, which can be quite costly (or impossible) to ensure. In the second place, linear epistemic relevance orderings of the propositions in a belief base make belief base contraction easy to implement by simply dropping the lowest ranked propositions in any conflicting set (as in rup [ McAllester, 1982 ] Moreover, this form of belief base contraction is fully rational, and corresponds to maxichoice contraction on belief bases [ Nebel, 1989, Lemma 17ff ] Representing epistemic entrenchment, however, can be more costly. Gardenfors and Makinson [ 1988 ] show that epistemic entrenchment orders ....

D. A. McAllester. Reasoning Utility Package user's manual. AIM 667, Massachusetts Institute of Technology, Artificial Intelligence Laboratory, Cambridge, MA, 1982.

....exhibited by maxichoice contraction of belief states. Moreover, maxichoice base contraction is just as rational as the more complicated base contraction. Since it can be iterated easily and does not introduce complicated disjunctions, it is to be preferred in practice. He points to McAllester s [31] rup system as an implementation of just this sort of rational belief base revision (ignoring the issue of rup s deductive incompleteness) 3.3 Epistemic relevance vs. epistemic entrenchment The orders of epistemic relevance and epistemic entrenchment are superficially different in that ....

....or y x, in contrast with the dominance condition (2) on epistemic entrenchment orderings, which can be quite costly (or impossible) to ensure. In the second place, linear epistemic relevance orderings of the propositions in a belief base make belief base contraction easy to implement (as in rup [31]) by simply dropping the lowest ranked propositions in any conflicting set. Moreover, this form of belief base contraction is fully rational, and corresponds to maxichoice contraction [32, Lemma 17ff] Epistemic entrenchment is not entirely impractical, as Gardenfors and Makinson [16] show that ....

D. A. McAllester. Reasoning Utility Package user's manual. AIM 667, Massachusetts Institute of Technology, Artificial Intelligence Laboratory, 545 Technology Square, Cambridge, MA, 02139, 1982.

....a polynomial algorithm for satisfiability (by restricting the language, for instance) then unambiguous base revision will be itself polynomial. This result gives a formal justification for the claim made in [ Nebel, 1989 ] that this form of revision is similar to the functionality the rup system [ McAllester, 1982 ] offers in an abstract sense, though. 10 The important point to note is that a feasible implementation of belief revision is possible if we restrict ourselves to polynomial methods for satisfiability by restricting the language or by tolerating incompleteness and by using a polynomial method ....

David A. McAllester. Reasoning utility package user's manual. AI Memo 667, AI Laboratory, Massachusetts Institute of Technology, Cambridge, Mass., April 1982.

....be briefly sketched. 2.1 CAKE cake is a the representation kernel of a programmer s apprentice system [52] One component is the plan calculus component, which reasons about programs represented by using a flow chart style formalism with pre and postconditions. The second component is rup [32] a TMS that supports reasoning in variablefree predicate logic with equality. The link between these two formalisms is established by a sort of translation between the two formalisms. Plans and programs used in the plan calculus component are mapped to assertions that are stored in the active ....

D. A. McAllester. Reasoning Utility Package User's Manual. AI Memo 667, AI Laboratory, Massachusetts Institute of Technology, Cambridge, Mass., Apr. 1982.

....calculus decision procedure [DaPu60] The same instances of clauses may appear in more than one case of a case analysis, although in each case there may be no duplication. Some early provers used duplication by case analysis, but few if any clause form provers currently use it. Andrews prover [MCA82] seems to be one of the few current provers that use duplication by case analysis. We now make more precise how resolution and other methods involve the duplication of instances. Given clauses C and D, resolution computes a most general unifier Theta and then combines literals of C Theta and ....

....solved and we know S is unsatisfiable. Otherwise, more rounds of hyper linking and so on, are done. The propositional calculus prover is a modification of the Davis Putnam procedure [DaPu60] It uses a technique similar to dependency directed backtracking [dK86a] dK89] and McAllester s RUP [McA80] McA82] in the area of Artificial Intelligence. One disadvantage of the Davis Putnam procedure is that it carries out two cases in a case analysis all the time, which results in unnecessary duplication by case analysis. The motivation of our propositional calculus prover is to remove such unnecessary ....

D. McAllester. Reasoning utility package user's manual. Technical Report AIM-667, Artificial Intelligence Laboratory, MIT, Cambridge, MA, 1982.

....difficulty of quickly computing this form of conservative updates, however, RMS only approximates conservative updates. Reason maintenance systems developed later use simpler notions of conservatism which may be computed more rapidly. For updating beliefs based on monotonic reasons, for example, McAllester s (1982; 1990) efficient so called boolean constraint propagation technique maintains orderings of the premises generating belief sets. As Nebel (1989; 1990) points out, this ordering is analogous to an entrenchment ordering, and permits rapid conservative updates. The belief revision operations of ....

McAllester, David 1982. Reasoning Utility Package user's manual. Artificial Intelligence Memo 667, Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA.

....that almost all representation formalisms used in Artficial Intelligence are intractable or even undecidable. Therefore in practical systems tractable but incomplete algorithms are often used, as for example, in the terminological component of KL TWO [9] in the reasoning maintenance system RUP [14], and in Allen s temporal reasoner [1] 10 If, however, completeness is a goal one cannot dispense with, expressive power has to be severely restricted. In our case, one solution would be to sacrifice all operators that state relationships between roles, i.e. primitive subrole introduction and ....

McAllester, D. A., Reasoning Utility Package User's Manual, AI Memo 667, AI Lab., MIT, Cambridge, Mass., 1982.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC