|
Title:
|
SEMANTIC TAGGING OF REQUIREMENT
DESCRIPTIONS: A TRANSFORMER-BASED APPROACH |
|
Author(s):
|
Joschka Kersting and Frederik S. Bäumer |
|
ISBN:
|
978-989-8704-24-5 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2020 |
|
Edition:
|
Single |
|
Keywords:
|
Software Requirements, Natural Language Processing, Transfer Learning, On-The-Fly Computing |
|
Type:
|
Full |
|
First Page:
|
119 |
|
Last Page:
|
123 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The idea of On-The-Fly Computing is to compose and provide software services ad hoc to end users. This is a highly
automated process, which has to start with requirement descriptions in natural language. The main reason is that end
users prefer natural language software requirements over formal language due to the usability and accessibility. However,
there are good reasons for not using natural language, such as the susceptibility to errors. Nevertheless, to enable
processing, a machine interpretation of these initial natural language text data is of fundamental importance. In this short
paper, we present our ongoing work on REaCT2, a tool that is able to process requirement documents regarding semantic
roles. This information can be reliably extracted and used to provide software services. More specifically, REaCT2
achieves an F1 Score of 90% for the semantic tagging of requirements and outperforms existing software solutions in this
task. |
|
|
|
|
|
|
