Papers › Contrastive Code Representation Learning

Contrastive Code Representation Learning

9 Jul 2020EMNLP 2021 11arXiv:2007.04973archive 2025-07-28

Paras Jain, Ajay Jain, Tianjun Zhang, Pieter Abbeel, Joseph E. Gonzalez, Ion Stoica

Recent work learns contextual representations of source code by reconstructing tokens from their context. For downstream semantic understanding tasks like summarizing code in English, these representations should ideally capture program functionality. However, we show that the popular reconstruction-based BERT model is sensitive to source code edits, even when the edits preserve semantics. We propose ContraCode: a contrastive pre-training task that learns code functionality, not form. ContraCode pre-trains a neural network to identify functionally similar variants of a program among many non-equivalent distractors. We scalably generate these variants using an automated source-to-source compiler as a form of data augmentation. Contrastive pre-training improves JavaScript summarization and TypeScript type inference accuracy by 2% to 13%. We also propose a new zero-shot JavaScript code clone detection dataset, showing that ContraCode is both more robust and semantically meaningful. On it, we outperform RoBERTa by 39% AUROC in an adversarial setting and up to 5% on natural code.

Paper PDF Conference PDF Code

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

parasj/contracode officialmentioned in paperpytorchApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Clone Detection Contrastive Learning Data Augmentation Method name prediction Representation Learning Source Code Summarization Type prediction

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
Method name prediction	CodeSearchNet	ContraCode	F1	17.24	#1 of 1	Archive leaderboard	report
Source Code Summarization	CodeSearchNet	ContraCode	F1	17.24	#1 of 1	Archive leaderboard	report
Type prediction	DeepTyper	ContraCode	Accuracy@5	84.60	#1 of 1	Archive leaderboard	report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Adam Attention Attention Dropout BERT Dense Connections Dropout Layer Normalization Linear Layer Linear Warmup With Linear Decay Multi-Head Attention Residual Connection RoBERTa Softmax Weight Decay WordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections