We discuss P#, our implementation of a tool which allows interoperation between a concurrent superset of the Prolog programming language and C#. This enables Prolog to be used as a native implementation language for Microsoft's .NET platform. P# compiles a linear logic extension of Prolog to C# source code. We can thus create C# objects from Prolog and use C#'s graphical, networking and other libraries. P# was developed from a modified port of the Prolog to Java translator, Prolog Cafe. We add language constructs on the Prolog side which allow concurrent Prolog code to be written. We add a primitive predicate which evaluates a Prolog structure on a newly forked thread. Communication between threads is based on the unification of variables contained in such a structure. It is also possible for threads to communicate through a globally accessible table. All of the new features are available to the programmer through new built-in Prolog predicates. We present three case studies. The first is an application which allows several users to modify a database. The users are able to disconnect from the database and to modify their own copies of the data before reconnecting. On reconnecting, conflicts must be resolved. The second is an object-oriented assistant, which allows the user to query the contents of a C# namespace or Java package. The third is a tool which allows a user to interact with a graphical display of the inheritance tree. Finally, we optimize P#'s runtime speed by translating some Prolog predicates into more idiomatic C# code than is produced by a naive port of Prolog Cafe. This is achieved by observing that semi-deterministic predicates (being those which always either fail or succeed with exactly one solution) that only call other semi-deterministic predicates enjoy relatively simple control flow. We make use of the fact that Prolog programs often contain predicates which operate as functions, and that such predicates are usually semi-deterministic.

by Hwang Hoi Yee Cothan. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1996. / Includes bibliographical references (leaves 68-70). / Abstract --- p.ii / Acknowledgement --- p.iii / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Objectives of the Thesis --- p.1 / Chapter 1.2 --- Overview of the Thesis --- p.7 / Chapter 2 --- Background and Related Work --- p.8 / Chapter 2.1 --- Deductive Databases --- p.8 / Chapter 2.2 --- Disjunctive Deductive Databases --- p.10 / Chapter 2.3 --- Model tree for disjunctive deductive databases --- p.11 / Chapter 3 --- Preliminary --- p.13 / Chapter 3.1 --- Disjunctive Logic Program --- p.13 / Chapter 3.2 --- Data-disjunctive Logic Program --- p.14 / Chapter 4 --- Semantics of Data-disjunctive Logic Program --- p.17 / Chapter 4.1 --- Model-theoretic semantics --- p.17 / Chapter 4.2 --- Fixpoint semantics --- p.20 / Chapter 4.2.1 --- Fixpoint operators corresponding to the MMSpDD --- p.22 / Chapter 4.2.2 --- "Fixpoint operator corresponding to the contingency model, CMP" --- p.25 / Chapter 4.3 --- Equivalence between the model-theoretic and fixpoint semantics --- p.26 / Chapter 4.4 --- Operational Semantics --- p.30 / Chapter 4.5 --- Correspondence with the I-table --- p.31 / Chapter 5 --- Disjunctive Deductive Databases --- p.33 / Chapter 5.1 --- Disjunctions in deductive databases --- p.33 / Chapter 5.2 --- Relation between predicates --- p.35 / Chapter 5.3 --- Transformation of Disjunctive Deductive Data-bases --- p.38 / Chapter 5.4 --- Query answering for Disjunctive Deductive Data-bases --- p.40 / Chapter 6 --- Magic for Data-disjunctive Deductive Database --- p.44 / Chapter 6.1 --- Magic for Relevant Answer Set --- p.44 / Chapter 6.1.1 --- Rule rewriting algorithm --- p.46 / Chapter 6.1.2 --- Bottom-up evaluation --- p.49 / Chapter 6.1.3 --- Examples --- p.49 / Chapter 6.1.4 --- Discussion on the rewriting algorithm --- p.52 / Chapter 6.2 --- Alternative algorithm for Traditional Answer Set --- p.54 / Chapter 6.2.1 --- Rule rewriting algorithm --- p.54 / Chapter 6.2.2 --- Examples --- p.55 / Chapter 6.3 --- Contingency answer set --- p.56 / Chapter 7 --- Experiments and Comparison --- p.57 / Chapter 7.1 --- Experimental Results --- p.57 / Chapter 7.1.1 --- Results for the Traditional answer set --- p.58 / Chapter 7.1.2 --- Results for the Relevant answer set --- p.61 / Chapter 7.2 --- Comparison with the evaluation method for Model tree --- p.63 / Chapter 8 --- Conclusions and Future Work --- p.66 / Bibliography --- p.68

Tese de mestrado integrado. Engenharia Informática e Computação. Faculdade de Engenharia. Universidade do Porto. 2010

This thesis examines a novel induction-based framework for logic programming. Limiting programs are logic programs distinguished by two features, in general they contain an infinite data stream over which induction will be performed, and in general it is not possible for a system to know when a solution for any program is correct. These facts are characteristic of some problems involving induction in artificial intelligence, and several problems in knowledge representation and logic programming have exactly these properties. This thesis presents a specification language for problems with an inductive nature, limiting programs, and a resolution based system, limiting resolution, for solving these problems. This framework has properties which guarantee that the system will converge upon a particular answer in the limit. Solutions to problems which have such an inductive property by nature can be implemented using the language, and solved with the solver. For instance, many classification problems are inductive by nature. Some generalized planning problems also have the inductive property. For a class of generalized planning problems, we show that identifying a collection of domains where a plan reaches a goal is equivalent to producing a plan. This thesis gives examples of both. Limiting resolution works by a generate-and-test strategy, creating a potential solution and iteratively looking for a contradiction with the growing stream of data provided. Limiting resolution can be implemented by modifying conventional PROLOG technology. The generateand- test strategy has some inherent inefficiencies. Two improvements have arisen from this work; the first is a tabling strategy which records previously failed attempts to produce a solution and thereby avoids redundant test steps. The second is based on the heuristic observation that for some problems the size of the test step is proportional to the closeness of the generated potential-solution to the real solution, in a suitable metric. The observation can be used to improve the performance of limiting resolution. Thus this thesis describes, from theoretical foundations to implementation, a coherent methodology for incorporating induction into existing general A.I. programming techniques, along with examples of how to perform such tasks.

published_or_final_version / Computer Science / Master / Master of Philosophy

Graduate

Constraint logic programming (CLP) techniques can be used in protein side-chain placement, an important sub-task in comparative modelling. In a simple formulation values for domain variables represent rotamer side-chain conformations, and constraints represent atomic clashes. These constraints can be visualised using a "rotamer contact map", and observations made with this visualisation tool have been used to develop a strategy that overcomes limitations present in CLP caused by over-constrained residues. Null rotamers provide a mechanism that can automatically identify over-constrained residues. The use of null rotamers makes possible an iterative modelling strategy where, at each iteration, a CLP program is generated automatically; each program representing successively tighter packing constraints corresponding to larger atomic radii. Different CLP enumeration heuristics have been evaluated for use with this side-chain placement method, and it has been tested with several different rotamer libraries; a backbone-dependent rotamer library, when used with first-fail enumeration heuristics, was shown to be the most successful. Side-chain conformations predicted by this CLP method compare favourably against those predicted using other side-chain placement methods. The CLP method has been applied to two modelling problems. The first involved building models of class II MHC molecules in order to increase the utility of a peptide threading program. This program uses an allele's known or modelled 3D structure with a heuristic scoring function to predict peptides that are likely to bind to it - thus using CLP to model class II MHC alleles increases the program's utility. The second application used the CLP method to build structures of ribosome inactivating proteins (RIPs). These models were built using CLP together with comparative modelling approaches, and a model of bouganin, a recently identified wild RIF protein, has been built to help design engineered therapeutic proteins.

No description available.

No description available.

No description available.