Learning to Represent Edits

Abstract

We introduce the problem of learning distributed representations of edits. By combining a "neural editor" with an "edit encoder", our models learn to represent the salient information of an edit and can be used to apply edits to new inputs. We experiment on natural language and source code edit data. Our evaluation yields promising results that suggest that our neural network models learn to capture the structure and semantics of edits. We hope that this interesting task and data source will inspire other researchers to work further on this problem.

arXiv commentICLR 2019

References in corpus (14)

• Neural Message Passing for Quantum Chemistry
• Effective Approaches to Attention-based Neural Machine Translation
• Gated Graph Sequence Neural Networks
• Constrained Graph Variational Autoencoders for Molecule Design
• Structured Generative Models of Natural Source Code
• A Retrieve-and-Edit Framework for Predicting Structured Outputs
• Differential Recurrent Neural Networks for Action Recognition
• Improved Variational Autoencoders for Text Modeling using Dilated Convolutions
• Tree-to-tree Neural Networks for Program Translation
• Generating Sentences by Editing Prototypes
• TRANX: A Transition-based Neural Abstract Syntax Parser for Semantic Parsing and Code Generation
• Learning Quick Fixes from Code Repositories
• Deep Differential Recurrent Neural Networks
• WikiAtomicEdits: A Multilingual Corpus of Wikipedia Edits for Modeling Language and Discourse

Cited by in corpus (21)

• Neural Transfer Learning for Repairing Security Vulnerabilities in C Code
• Graph Neural Networks Meet Neural-Symbolic Computing: A Survey and Perspective
• CODIT: Code Editing with Tree-Based Neural Models
• Graph Neural Networks for Natural Language Processing: A Survey
• A Literature Study of Embeddings on Source Code
• Commit2Vec: Learning Distributed Representations of Code Changes
• Editing-Based SQL Query Generation for Cross-Domain Context-Dependent Questions
• ProGraML: Graph-based Deep Learning for Program Optimization and Analysis
• Learning Sparse Prototypes for Text Generation
• Neural Bug Finding: A Study of Opportunities and Challenges
• Fact-based Text Editing
• Graph-to-Tree Neural Networks for Learning Structured Input-Output Translation with Applications to Semantic Parsing and Math Word Problem
• Learning the Relation between Code Features and Code Transforms with Structured Prediction
• Learning to Represent Programs with Heterogeneous Graphs
• Parallel Iterative Edit Models for Local Sequence Transduction
• Variational Inference for Learning Representations of Natural Language Edits
• Learning Non-Monotonic Automatic Post-Editing of Translations from Human Orderings
• Assessing the Effectiveness of Syntactic Structure to Learn Code Edit Representations
• Deep Data Flow Analysis
• World Programs for Model-Based Learning and Planning in Compositional State and Action Spaces
• An Edit-centric Approach for Wikipedia Article Quality Assessment