This paper describes a modified learning classifier system for evolving grammars from small, handcrafted grammars. The algorithm uses distinct discovery and evaluation steps, and a culling step to reduce ambiguity. It performs surprisingly well in developing grammars for very complex sentences in a restricted domain of English, and in some experiments the grammar's competence improved up to tenfold.

System-level design requires behavioral models of its complex components in order to validate the designs, synthesize implementations and generate tests. Descriptions of these components are often only available as English descriptions. Developing behavioral models from natural language descriptions has been a tedious and time consuming task. The paper describes the semiautomatic generation of conceptual models from text and a simulator for these models. The conceptual modeling language is a knowledge representation notation (semantic hypergraphs), and simulation of these abstract descriptions based on the VHDL execution model is described. Simulation at the conceptual level avoids having to add the extraneous declarations and other syntactic artifacts necessary with more conventional description languages. These declarations not only take time, but also prejudice the implementation.

Partitioning is a very important task in hardware/software co-design. Generally the size of the edge cut-set is used to evaluate the communication cost. When communication between components is through buffered channels, the size of the edge cut-set is not adequate to estimate the buffer size. A second important factor to measure the quality of partitioning is the system delay. Most partitioning approaches use the number of nodes/functions in each partition as constraints and attempt to minimize the communication cost. The data dependencies among nodes/functions, and their delays are not considered. In this paper we present partitioning with two objectives: (1) buffer size, which is estimated by analyzing the data flow patterns of the CDFG, and solved as a clique partitioning problem, and (2) the system delay that is estimated using List Scheduling. We pose the problem as a combinatorial optimization and use an efficient non-deterministic search algorithm called Problem-Space Genetic Algorithm to search for the optimum. Results are compared with those produced by simulated annealing.

Initial requirements for new digital systems and products that are generally expressed in a variety of notations including diagrams and natural language can be automatically translated to a common knowledge representation for integration, for consistency and completeness analysis, and for further automatic synthesis. Block diagrams, flowcharts, timing diagrams, and English as used in specifying digital systems requirements are considered as examples of source notations for system requirements. The knowledge representation selected for this work is a form of semantic network called conceptual graphs. For each source notation, a basis set or semantic primitives in terms of conceptual graphs is given, together with an algorithm for automatically generating conceptual structures from the notation. The automatic generation of conceptual structures from English presumes a restricted sublanguage of English and feedback to the author for verification of the interpretation. Mechanisms for integrating the separate conceptual structures generated from individual requirements expressions using schemata are discussed, and methods are illustrated for consistency and completeness analysis.

The development of a restricted natural language for expressing requirements is described. the purpose of the language is to support readable specifications, which can be analyzed for errors and automatically interpreted. the specific language designed here is for specification of digital systems. Semantics are represented in a type of semantic network constructed of concepts and relations. the semantic basis, consisting of concept and relation types together with semantic patterns, is developed from an examination of several formal specification and design notations, together with a selection of natural language statements from product descriptions. the syntax of the language is developed from a syntactic analysis of natural language statements, and the productions of the grammar are selected to maximize coverage of syntactic structures used in the selection. © 1995 John Wiley & Sons, Inc.

Addresses the problem of comparing and unifying temporal relationships between activities expressed in timing diagrams and natural language narrative (English). This problem often occurs in specifications expressing behavioral requirements and constraints. The approach followed is to translate both diagrams and text into a common knowledge representation (conceptual graphs) employing temporal relations developed for temporal interval logic. In this knowledge representation, the requirements may be integrated, checked for inconsistencies, and subjected to additional temporal reasoning. An algorithm of polynomial complexity for generating a compact representation of temporal relationships from timing diagrams is presented. Generation of comparable conceptual graphs from English statements is described by using examples. Integrating conceptual graphs from timing diagrams and sentences while checking for inconsistencies is also of polynomial complexity.>

The requirements and specifications documents which initiate and control design and development projects typically use a variety of formal and informal notational systems. The goal of the research reported is to automatically interpret requirement documents expressed in a variety of notations and to integrate the interpretations in order to support requirements analysis and synthesis from them. Because the source notations include natural language, a form of semantic net called conceptual graphs is adopted as the intermediate knowledge representation for expressing interpretations and integrating them. The focus is to describe the interpretation or mapping of a few requirements notations to conceptual graphs, and to indicate the process of joining these interpretations.>

The preparation of a valid specification is a vital but time consuming step in the design cycle. The purpose of the project described is to significantly reduce the time to produce specifications through the development of applicable design automation tools. Specifications are typically expressed in a variety of informal or ad hoc notations, including natural language. The specification requirements must be interpreted, integrated and checked after which an engineering model or a prototype can be synthesized. The author describes an approach to computer-aided support for these tasks by mapping requirements in the various notations onto a common knowledge representation notation in which the requirements may be integrated and checked for consistency and completeness.>

Reports on research to design a Specification Interpreter for automatically extracting engineering models from English language specifications. The role of the Specification Interpreter in Rapid System Prototyping is suggested. The engineering models extracted from the English Specifications will be expressed both in a graphical form for feedback to the specification author (or 'customer') and in a formal design representation such as those used as input to behavioral and logic synthesis. While a great deal of work remains to he done, a design strategy for an automated interpreter has been developed and part of the system has reached the prototype stage.>

This paper reports on a research project to design a system for automatically interpreting English specifications of digital systems in terms of design representation formalisms currently employed in CAD systems. The necessary processes involve the machine analysis of English and the synthesis of models from the specifications. The approach being investigated is interactive and consists of syntactic scanning, semantic analysis, interpretation generation, and model integration.