We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.SE

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Software Engineering

Title: HAConvGNN: Hierarchical Attention Based Convolutional Graph Neural Network for Code Documentation Generation in Jupyter Notebooks

Abstract: Jupyter notebook allows data scientists to write machine learning code together with its documentation in cells. In this paper, we propose a new task of code documentation generation (CDG) for computational notebooks. In contrast to the previous CDG tasks which focus on generating documentation for single code snippets, in a computational notebook, one documentation in a markdown cell often corresponds to multiple code cells, and these code cells have an inherent structure. We proposed a new model (HAConvGNN) that uses a hierarchical attention mechanism to consider the relevant code cells and the relevant code tokens information when generating the documentation. Tested on a new corpus constructed from well-documented Kaggle notebooks, we show that our model outperforms other baseline models.
Subjects: Software Engineering (cs.SE); Machine Learning (cs.LG)
Cite as: arXiv:2104.01002 [cs.SE]
  (or arXiv:2104.01002v2 [cs.SE] for this version)

Submission history

From: Dakuo Wang [view email]
[v1] Wed, 31 Mar 2021 22:36:41 GMT (1519kb,D)
[v2] Wed, 8 Sep 2021 18:49:12 GMT (7309kb,D)

Link back to: arXiv, form interface, contact.