Learning Sequential and Structural Information for Source Code Summarization

Choi, Young-Dae; Bak, JinYeong; Na, CheolWon; Lee, Jee-Hyong

doi:10.18653/v1/2021.findings-acl.251

Cited by 19 publications

(19 citation statements)

References 14 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, the power of the corresponding code-to-text models may be affected. To this end, more and more researchers have started to train the learning models with code's structural information, which can be obtained from the code's AST 16,17,[19][20][21][22]25,[31][32][33][34][35] and the Application Programming Interface (API) sequence. 13,36,37 For example, Wan et al 21 put forward a framework based on reinforcement learning to automatically summarize code snippets.…”

Section: Related Workmentioning

confidence: 99%

PyScribe–Learning to describe python code

Guo,

Liu,

Liu

et al. 2023

Softw Pract Exp

View full text Add to dashboard Cite

Code comment generation, which attempts to summarize the functionality of source code in textual descriptions, plays an important role in automatic software development research. Currently, several structural neural networks have been exploited to preserve the syntax structure of source code based on abstract syntax trees (ASTs). However, they can not well capture both the long‐distance and local relations between nodes while retaining the overall structural information of AST. To mitigate this problem, we present a prototype tool titled PyScribe, which extends the Transformer model to a new encoder‐decoder‐based framework. Particularly, the triplet position is designed and integrated into the node‐level and edge‐level structural features of AST for producing Python code comments automatically. This paper, to the best of our knowledge, makes the first effort to model the edges of AST as an explicit component for improved code representation. By specifying triplet positions for each node and edge, the overall structural information can be well preserved in the learning process. Moreover, the captured node and edge features go through a two‐stage decoding process to yield higher qualified comments. To evaluate the effectiveness of PyScribe, we resort to a large dataset of code‐comment pairs by mining Jupyter Notebooks from GitHub, for which we have made it publicly available to support further studies. The experimental results reveal that PyScribe is indeed effective, outperforming the state‐ofthe‐art by achieving an average BLEU score (i.e., av‐BLEU) of 0.28.

show abstract

Section: Related Workmentioning

confidence: 99%

PyScribe–Learning to describe python code

Guo,

Liu,

Liu

et al. 2023

Softw Pract Exp

View full text Add to dashboard Cite

show abstract

“…The encodings are then aggregated for generating the summary. Choi et al [11] apply graph convolutions to obtain node representations from an AST and then input the sequence of node representations into the Transformer layers for code summarization. Wu et al [44] construct a multi-view adjacent matrix to represent the relationships between the tokens in source code, and use it to guide the self-attention computation in Transformer.…”

Section: Related Workmentioning

confidence: 99%

“…The encodings are then aggregated for generating the summary. • mAST+GCN [11]. The model applies graph convolutions to obtain node representations from an AST and then inputs the sequence of node representations into the Transformer layers for code summarization.…”

Section: The Comparative Modelsmentioning

confidence: 99%

GypSum: Learning Hybrid Representations for Code Summarization

Wang,

Dong,

et al. 2022

Preprint

View full text Add to dashboard Cite

Code summarization with deep learning has been widely studied in recent years. Current deep learning models for code summarization generally follow the principle in neural machine translation and adopt the encoder-decoder framework, where the encoder learns the semantic representations from source code and the decoder transforms the learnt representations into human-readable text that describes the functionality of code snippets. Despite they achieve the new state-of-the-art performance, we notice that current models often either generate less fluent summaries, or fail to capture the core functionality, since they usually focus on a single type of code representations. As such we propose GypSum, a new deep learning model that learns hybrid representations using graph attention neural networks and a pre-trained programming and natural language model. We introduce particular edges related to the control flow of a code snippet into the abstract syntax tree for graph construction, and design two encoders to learn from the graph and the token sequence of source code, respectively. We modify the encoder-decoder sublayer in the Transformer's decoder to fuse the representations and propose a dual-copy mechanism to facilitate summary generation. Experimental results demonstrate the superior performance of GypSum over existing code summarization models. CCS CONCEPTS• Computing methodologies → Neural networks; • Software and its engineering → Documentation.

show abstract

“…Past works handling programs and code have focused on enriching their models with incorporating more semantic and syntactic information from code [1,12,33,43]. Some prior works have cast the SCS as a sequence classification task, where the code is represented as a textual sequence and input pair (q i , c j ) is concatenated with a special separator symbol into a single sequence, and the output is the score rij : rij = f (q i , c j ).…”

Section: Neural Models For Semantic Code Searchmentioning

confidence: 99%

NS3: Neuro-Symbolic Semantic Code Search

Arakelyan¹,

Hakhverdyan²,

Allamanis³

et al. 2022

Preprint

View full text Add to dashboard Cite

Semantic code search is the task of retrieving a code snippet given a textual description of its functionality. Recent work has been focused on using similarity metrics between neural embeddings of text and code. However, current language models are known to struggle with longer, compositional text, and multi-step reasoning.To overcome this limitation, we propose supplementing the query sentence with a layout of its semantic structure. The semantic layout is used to break down the final reasoning decision into a series of lower-level decisions. We use a Neural Module Network architecture to implement this idea. We compare our model -NS 3 (Neuro-Symbolic Semantic Search) -to a number of baselines, including state-of-the-art semantic code retrieval methods, and evaluate on two datasets -CodeSearchNet and Code Search and Question Answering. We demonstrate that our approach results in more precise code retrieval, and we study the effectiveness of our modular design when handling compositional queries 1 .Preprint. Under review.

show abstract

Learning Sequential and Structural Information for Source Code Summarization

Cited by 19 publications

References 14 publications

PyScribe–Learning to describe python code

PyScribe–Learning to describe python code

GypSum: Learning Hybrid Representations for Code Summarization

NS3: Neuro-Symbolic Semantic Code Search

Contact Info

Product

Resources

About