Source code summarization aims to generate concise descriptions for code snippets in a natural language, thereby facilitates program comprehension and software maintenance. In this paper, we propose a novel approach– GSCS –to automatically generate summaries for Java methods, which leverages both semantic and structural information of the code snippets. To this end, GSCS utilizes Graph Attention Networks to process the tokenized abstract syntax tree of the program, which employ a multi-head attention mechanism to learn node features in diverse representation sub-spaces, and aggregate features by assigning different weights to its neighbor nodes. GSCS further harnesses an additional RNN-based sequence model to obtain the semantic features and optimizes the structure by combining its output with a transformed embedding layer. We evaluate our approach on two widely-adopted Java datasets; the experiment results confirm that GSCS outperforms the state-of-the-art baselines.

Deep neural networks (DNNs) have shown remarkable performance in a variety of domains such as computer vision, speech recognition, and natural language processing. Recently they also have been applied to various software engineering tasks, typically involving processing source code. DNNs are well-known to be vulnerable to adversarial examples, i.e., fabricated inputs that could lead to various misbehaviors of the DNN model while being perceived as benign by humans. In this paper, we focus on the code comment generation task in software engineering and study the robustness issue of the DNNs when they are applied to this task. We propose ACCENT (Adversarial Code Comment gENeraTor) , an identifier substitution approach to craft adversarial code snippets, which are syntactically correct and semantically close to the original code snippet, but may mislead the DNNs to produce completely irrelevant code comments. In order to improve the robustness, ACCENT also incorporates a novel training method, which can be applied to existing code comment generation models. We conduct comprehensive experiments to evaluate our approach by attacking the mainstream encoder-decoder architectures on two large-scale publicly available datasets. The results show that ACCENT efficiently produces stable attacks with functionality-preserving adversarial examples, and the generated examples have better transferability compared with the baselines. We also confirm, via experiments, the effectiveness in improving model robustness with our training method.

Code summarization aims to automatically generate functionality descriptions of code snippets. Faithful metrics are needed to measure to which degree the machine generated summaries capture the semantics of the code snippets. Most commonly used metrics in code summarization, such as BLEU -4, METEOR, and ROUGE-L, originate from machine translation and text summarization, and have constantly been found to be inconsistent with human assessment. In this paper, we propose a novel evaluation metric, Consensus-based Code Summarization Evaluation (CCSE), which assigns different semantic weights to the n-grams of the summary. We also provide an algorithm to match the n-gram pairs from the reference and candidate based on the similarities. To validate the effectiveness of our proposed metric, we collect summary pairs from two public Java datasets and calculate the correlation coefficients between CCSE and the human evaluations. The experiment results show that, compared with BLEU-4, METEOR, and ROUGE-L, CCSE is more consistent with the scores assessed by human developers.