Open AccessProgram Semantics-based Task Decomposition
Author(s) -
Zhiming Wang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1684/1/012051
Subject(s) - computer science , task (project management) , semantics (computer science) , decomposition , set (abstract data type) , natural language processing , programming language , artificial intelligence , code (set theory) , software , engineering , ecology , systems engineering , biology
Recently, automatic code generation technology has been extensively studied and gradually becomes an important part of software development. However, when facing coarsegrained, highly general task description, the current automatic code generation technology still cannot achieve good results. Therefore, we propose program semantics-based task decomposition technology. For a task description given by the user, natural language processing and deep learning technology are used to learn and understand it. Combined with the program semantics contained therein, we decompose the task into ordered subtask sequence. We evaluate our proposed model on constructed data set and achieve a BLEU-4 score of 18.13.