Code comments are crucial in software engineering, aiding in program maintenance and code reuse. The process of generating clear and descriptive code comments, outlining code functionality, is called code summarizatio...Code comments are crucial in software engineering, aiding in program maintenance and code reuse. The process of generating clear and descriptive code comments, outlining code functionality, is called code summarization. Existing code summarization methods are typically trained using transformer-based models. However, these trained models often possess limited parameters and lack specific training tasks, hindering their ability to capture code semantics effectively. This paper uses a high-capacity pre-trained model, CodeT5, for code summarization. CodeT5 is designed with an encoder-decoder architecture that excels in code summarization tasks. Furthermore, we adopt a novel paradigm, "pre-train, prompt, predict", to unlock the knowledge embedded within CodeT5. We devise a prompt template to convert input code into code prompts and fine-tune CodeT5 with these prompts—a process we term prompt tuning. Our effectiveness experiments demonstrate that prompt tuning CodeT5 with only 40% of the dataset can achieve comparable performance to fine-tuning CodeT5 with 100% of the dataset. This means our approach is applicable in few-shot learning scenarios. Additionally, our prompt learning method is not sensitive to the size of the tuning dataset. Our practicality experiments show that the performance of prompt-tuned CodeT5 far surpasses that of transformer-based models trained on code-comment datasets collected from Stack Overflow.展开更多
Recent studies have applied different approaches for summarizing software artifacts, and yet very few efforts have been made in summarizing the source code fragments available on web. This paper investigates the feasi...Recent studies have applied different approaches for summarizing software artifacts, and yet very few efforts have been made in summarizing the source code fragments available on web. This paper investigates the feasibility of generating code fragment summaries by using supervised learning algorithms. We hire a crowd of ten individuals from the same work place to extract source code features on a cor- pus of 127 code fragments retrieved from Eclipse and Net- Beans Official frequently asked questions （FAQs）. Human an- notators suggest summary lines. Our machine learning algo- rithms produce better results with the precision of 82% and perform statistically better than existing code fragment classi- fiers. Evaluation of algorithms on several statistical measures endorses our result. This result is promising when employing mechanisms such as data-driven crowd enlistment improve the efficacy of existing code fragment classifiers.展开更多
This paper presents a literature review in the field of summarizing software artifacts, focusing on bug reports, source code, mailing lists and developer discussions artifacts. From Jan. 2010 to Apr. 2016, numerous su...This paper presents a literature review in the field of summarizing software artifacts, focusing on bug reports, source code, mailing lists and developer discussions artifacts. From Jan. 2010 to Apr. 2016, numerous summarization techniques, approaches, and tools have been proposed to satisfy the ongoing demand of improving software performance and quality and facilitating developers in understanding the problems at hand. Since aforementioned artifacts contain both structured and unstructured data at the same time, researchers have applied different machine learning and data mining techniques to generate summaries. Therefore, this paper first intends to provide a general perspective on the state of the art, describing the type of artifacts, approaches for summarization, as well as the common portions of experimental procedures shared among these artifacts. Moreover, we discuss the applications of summarization, i.e., what tasks at hand have been achieved through summarization. Next, this paper presents tools that are generated for summarization tasks or employed during summarization tasks. In addition, we present different summarization evaluation methods employed in selected studies as well as other important factors that are used for the evaluation of generated summaries such as adequacy and quality. Moreover, we briefly present modern communication channels and complementarities with commonalities among different software artifacts. Finally, some thoughts about the challenges applicable to the existing studies in general as well as future research directions are also discussed. The survey of existing studies will allow future researchers to have a wide and useful background knowledge on the main and important aspects of this research field.展开更多
文摘Code comments are crucial in software engineering, aiding in program maintenance and code reuse. The process of generating clear and descriptive code comments, outlining code functionality, is called code summarization. Existing code summarization methods are typically trained using transformer-based models. However, these trained models often possess limited parameters and lack specific training tasks, hindering their ability to capture code semantics effectively. This paper uses a high-capacity pre-trained model, CodeT5, for code summarization. CodeT5 is designed with an encoder-decoder architecture that excels in code summarization tasks. Furthermore, we adopt a novel paradigm, "pre-train, prompt, predict", to unlock the knowledge embedded within CodeT5. We devise a prompt template to convert input code into code prompts and fine-tune CodeT5 with these prompts—a process we term prompt tuning. Our effectiveness experiments demonstrate that prompt tuning CodeT5 with only 40% of the dataset can achieve comparable performance to fine-tuning CodeT5 with 100% of the dataset. This means our approach is applicable in few-shot learning scenarios. Additionally, our prompt learning method is not sensitive to the size of the tuning dataset. Our practicality experiments show that the performance of prompt-tuned CodeT5 far surpasses that of transformer-based models trained on code-comment datasets collected from Stack Overflow.
基金We would like to extend our gratitude to the individu- als who dedicated their time and effort to participate in crowdsourcing activ- ity and annotation of our code fragment corpus. This work was supported in part by National Program on Key Basic Research Project （2013CB035906）, in part by the New Century Excellent Talents in University （NCET-13-0073）, and in part by the National Natural Science Foundation of China （Grant Nos. 61175062, 61370144）.
文摘Recent studies have applied different approaches for summarizing software artifacts, and yet very few efforts have been made in summarizing the source code fragments available on web. This paper investigates the feasibility of generating code fragment summaries by using supervised learning algorithms. We hire a crowd of ten individuals from the same work place to extract source code features on a cor- pus of 127 code fragments retrieved from Eclipse and Net- Beans Official frequently asked questions （FAQs）. Human an- notators suggest summary lines. Our machine learning algo- rithms produce better results with the precision of 82% and perform statistically better than existing code fragment classi- fiers. Evaluation of algorithms on several statistical measures endorses our result. This result is promising when employing mechanisms such as data-driven crowd enlistment improve the efficacy of existing code fragment classifiers.
基金This work was supported in part by the National Basic Research 973 Program of China under Grant No. 2013CB035906, the Fundamental Research Funds for the Central Universities of China under Grant No. DUT13RC（3）53, and in part by the New Century Excellent Talents in University of China under Grant No. NCET-13-0073 and the National Natural Science Foundation of China under Grant No. 61300017.
文摘This paper presents a literature review in the field of summarizing software artifacts, focusing on bug reports, source code, mailing lists and developer discussions artifacts. From Jan. 2010 to Apr. 2016, numerous summarization techniques, approaches, and tools have been proposed to satisfy the ongoing demand of improving software performance and quality and facilitating developers in understanding the problems at hand. Since aforementioned artifacts contain both structured and unstructured data at the same time, researchers have applied different machine learning and data mining techniques to generate summaries. Therefore, this paper first intends to provide a general perspective on the state of the art, describing the type of artifacts, approaches for summarization, as well as the common portions of experimental procedures shared among these artifacts. Moreover, we discuss the applications of summarization, i.e., what tasks at hand have been achieved through summarization. Next, this paper presents tools that are generated for summarization tasks or employed during summarization tasks. In addition, we present different summarization evaluation methods employed in selected studies as well as other important factors that are used for the evaluation of generated summaries such as adequacy and quality. Moreover, we briefly present modern communication channels and complementarities with commonalities among different software artifacts. Finally, some thoughts about the challenges applicable to the existing studies in general as well as future research directions are also discussed. The survey of existing studies will allow future researchers to have a wide and useful background knowledge on the main and important aspects of this research field.
