Building specification structured rule automatic generation device and method

Abstract

The invention discloses a semantic parsing-based structured building specification automatic generation device, which comprises a coding module of building specification domain features, a generationmodule with SNL grammatical constraints and a two-stage generation module, and is characterized in that the coding module of the building specification domain features obtains term semantic vectors according to natural language description; the generation module with the SNL grammatical constraint realizes a generation function with the grammatical constraint; the generation module with the SNL grammatical constraint converts the generation process of the SNL into a selection task of a sentence pattern template and a filling task of sentence pattern details by utilizing the grammatical construction rule of the SNL and the detachable characteristic of the SNL, so that the limitation on the generation process of the decoder is realized; the two-stage generation module provides a function ofconverting a natural language rule into SNL; the two-stage generation module fuses an encoding method and a generation method with grammatical constraints to obtain a two-stage generation model so asto reduce errors caused by overfitting of a single model.

Description

technical field [0001] The present invention relates to a method for automatic generation of structured building codes, in particular to a device and method for realizing the conversion from natural language to structured building codes using semantic analysis technology, which is applied to the structured writing stage of building codes and the construction of building information models Review, design phase. Background technique [0002] As the level of digitization in the field of architectural design continues to deepen, the automatic review technology of architectural drawings is becoming more and more important. Among automatic review technologies, structured specifications are the basic data support. This is because the rules described in natural language are not enforceable. At this stage, the structure of current building codes is mainly written manually. This method mainly has the problems of long development cycle, high personnel requirements, and high collabor...

AI Technical Summary

Problems solved by technology

[0009] (1) Among the existing semantic analysis tasks, SQL and code generation are more common, and the methods include sketch filling and reinforcement learning methods, but they are mainly designed for SQL / Python, and the grammatical structure of these languages ​​​​is quite different from SNL

[0010] (2) The existing semantic analysis technology relies less on the background knowledge of a specific field, and there are a large number of proper nouns in the construction field, which need to be distinguished, otherwise the generated results will be poor

[0011] (3) Existing constrained generation techniques are often designed for deep recursive structures, which do not match the actual situation of SNL

If you forcibly support a deeper structure, there will be a problem of under-learning

Images