Abstract:In this paper, we describe an approach to translating the written content of the Korean Building Act into a computer-executable format for the purpose of evaluating building permit requirements. Among the various applications of building information modeling (BIM), we focus on automated design assessment and its rule-making process, which has been performed by both architects and software developers. Compared to the conventional rule-making approach, which is integrated with rule-checking software, our suggested mechanism (KBimLogic) of converting rule sentences from the Korean Building Act into computer-executable code (KBimCode) is a software-independent approach that separates the rule-making and rule-checking processes. We use the rule-making approach to translate the Korean Building Act into an explicit code that focuses on building permit requirements.Building permit-related regulations are defined in the Korean Building Act, but some of the sentences are ambiguous, and some implicit definitions hinder translation into an explicitly defined computer-executable form. Some building permit-related requirements vary by building type, administrative district, permitted date, and as-built date; thus, it is critical to design computable rules independent of specific proprietary software. The building permit system in Korea changes (as it does in other countries), making it critical that code compliance rules be kept up-to-date. Our work is motivated by such fundamentals and suggests a logic rule-based mechanism for use by non-programmers and a user-friendly approach to the rule-making process.The scope and major components of our research are as follows: (1) noun phrases classify building objects and associated properties from the Korean Building Act; (2) verb phrases derive high-level methods to construct the actual rule-checking body; and (3) the logic rule-based mechanism processes natural language sentences. In the actual implementation for translation, this approach has been demonstrated by each database and a GUI-based application named KBimLogic, for generating KBimCode. The KBimCode, which contains a set of specific building permit requirements translated from sentences, can be imported into our code-checking software (KBimAssess) using a specific file format.

Deep learning-based extraction of predicate-argument structure (PAS) in building design rule sentences☆

Integrating NLP and context-free grammar for complex rule interpretation towards automated compliance checking

Knowledge-informed semantic alignment and rule interpretation for automated compliance checking

Automated Building Information Modeling Compliance Check Through a Large Language Model Combined with Deep Learning and Ontology

Reasoning Makes Good Annotators : an Automatic Task-specific Rules Distilling Framework for Low-resource Relation Extraction

Translating building legislation into a computer-executable format for evaluating building permit requirements

Semiautomatic Rule Assist Architecture Modeling

DeepFacade: A Deep Learning Approach to Facade Parsing with Symmetric Loss.

A Building Regulation Question Answering System: A Deep Learning Methodology.

NLP-assisted Information Extraction for Layout Criteria of Substation

Rule-based information extraction for mechanical-electrical-plumbing-specific semantic web

Using Large Language Models for the Interpretation of Building Regulations

Extracting Domain Knowledge Elements of Construction Safety Management: Rule-Based Approach Using Chinese Natural Language Processing

CODE-ACCORD: A Corpus of Building Regulatory Data for Rule Generation towards Automatic Compliance Checking

Interpreting Deep Learning Model Using Rule-based Method

An Auto-Extraction Framework for CEP Rules Based on the Two-layer LSTM Attention Mechanism: A Case Study on City Air Pollution Forecasting

Extraction of Construction Quality Requirements from Textual Specifications via Natural Language Processing

Construction of design requirements knowledgebase from unstructured design guidelines using natural language processing

Research on Deep Learning-based AI Information Extraction Methods of Substation Engineering Design

LoCSGN: Logic-Contrast Semantic Graph Network for Machine Reading Comprehension

Embedded Machine Learning Systems For Robust Spoken Language Parsing