ProCQA: A Large-scale Community-based Programming Question Answering Dataset for Code Search
This addresses the need for better code search tools for developers by providing a more effective dataset and method, though it is incremental as it builds on existing pre-training techniques.
The paper tackles the problem of retrieval-based code question answering by introducing ProCQA, a large-scale dataset from StackOverflow, and a modality-agnostic contrastive pre-training approach, resulting in significant performance improvements across code retrieval benchmarks.
Retrieval-based code question answering seeks to match user queries in natural language to relevant code snippets. Previous approaches typically rely on pretraining models using crafted bi-modal and uni-modal datasets to align text and code representations. In this paper, we introduce ProCQA, a large-scale programming question answering dataset extracted from the StackOverflow community, offering naturally structured mixed-modal QA pairs. To validate its effectiveness, we propose a modality-agnostic contrastive pre-training approach to improve the alignment of text and code representations of current code language models. Compared to previous models that primarily employ bimodal and unimodal pairs extracted from CodeSearchNet for pre-training, our model exhibits significant performance improvements across a wide range of code retrieval benchmarks.