Reading between the Lines: Information Extraction from Industry Requirements
20037Authors
nan
Abstract
AbstractIndustry requirements describe the qualities that a project or a service must provide. Most requirements are, however, only available in natural language format and are embedded in textual documents. To be machine-understandable, a requirement needs to be represented in a logical format. We consider that a requirement consists of a scope, which is the requirement’s subject matter, a condition, which is any condition that must be fulfilled for the requirement to be relevant, and a demand, which is what is required. We introduce a novel task, the identification of the semantic components scope, condition, and demand in a requirement sentence, and establish baselines using sequence labelling and few-shot learning. One major challenge with this task is the implicit nature of the scope, often not stated in the sentence. By including document context information, we improved the average performance for scope detection. Our study provides insights into the difficulty of machine understanding of industry requirements and suggests strategies for addressing this challenge.
Keywords
Full Content
Proceedings of Recent Advances in Natural Language Processing, pages 703–711 Varna, Sep 4–6, 2023 https://doi.org/10.26615/978-954-452-092-2_076 703 Reading Between the Lines: Information Extraction from Industry Requirements Ole Magnus Holter University of Oslo / Norway olemholt@ifi.uio.no Basil Ell University of Oslo / Norway Bielefeld University / Germany basile@ifi.uio.no Abstract Industry requirements describe the qualities that a project or a service must provide. Most requirements are, however, only available in natural language format and are embed- ded in textual documents. To be machine- understandable, a requirement needs to be rep- resented in a logical format. We consider that a requirement consists of a scope, which is the requirement’s subject matter, a condition, which is any condition that must be fulfilled for the requirement to be relevant, and a demand, which is what is required. We introduce a novel task, the identification of the semantic com- ponents scope, condition, and demand in a requirement sentence, and establish baselines using sequence labelling and few-shot learning. One major challenge with this task is the im- plicit nature of the scope, often not stated in the sentence. By including document context information, we improved the average perfor- mance for scope detection. Our study provides insights into the difficulty of machine under- standing of industry requirements and suggests strategies for addressing this challenge. 1 Introduction Requirements are a critical part of the development process for products and services. They are docu- mented descriptions of the physical or functional qualities that a product or a service must have. In industry, requirements serve as a means of commu- nication between contractors and manufacturers, defining what is expected to be built and the quality standards to be met. Governments and interna- tional organizations may also impose requirements to ensure compliance with rules, regulations and standards. Requirements are included as part of the contract between two parties, making adherence to them a legal obligation. Violating the requirements can lead to legal implications and financial losses, underscoring the importance of careful specifica- tion and adherence to requirements throughout the development process. A requirement is typically associated with a spe- cific piece of equipment that needs to be built which is referred to as the scope of the requirement. How- ever, a requirement may only be relevant if certain conditions are met, which will be referred to as condition. The demand of the requirement is a feature or quality the scope must possess. As an example, consider the requirement equipment with a weight of more than 1000 kg shall have a weight certificate. Here, the scope is equipment, while the condition is with a weight of more than 1000 kg, and the demand is to have a weight certificate. Most requirements are expressed as natural lan- guage text and are embedded in documents. These documents often have a hierarchical structure with chapters, sections, and other headings, which pro- vide important context for understanding the re- quirements. When the number of requirements documented in this way increases, managing and maintaining these documents becomes a significant …
Content truncated to 500 words
Labeled Text
AbstractIndustry requirements describe the qualities that a project or a service must provide. Most requirements are, however, only available in natural language format and are embedded in textual documents. To be machine-understandable, a requirement needs to be represented in a logical format. We consider that a requirement consists of a scope, which is the requirement’s subject matter, a condition, which is any condition that must be fulfilled for the requirement to be relevant, and a demand, which is what is required. We introduce a novel task, the identification of the semantic components scope, condition, and demand in a requirement sentence, and establish baselines using sequence labelling and few-shot learning. One major challenge with this task is the implicit nature of the scope, often not stated in the sentence. By including document context information, we improved the average performance for scope detection. Our study provides insights into the difficulty of machine understanding of industry requirements and suggests strategies for addressing this challenge.
Publication Details
- Paper ID:
- 20037
- Year:
- 2023
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan