Cross-Domain Generalization Through Memorization: A Study of Nearest Neighbors in Neural Duplicate Question Detection

TL;DR

This paper investigates how neural representations and nearest neighbor methods can improve cross-domain duplicate question detection, showing robustness and outperforming traditional classification in various datasets.

Contribution

It introduces a k-nearest neighbor retrieval approach using neural embeddings for cross-domain DQD, demonstrating improved generalization over standard classifiers.

Findings

Robust cross-domain performance across multiple datasets.

Outperforms cross-entropy classification in several scenarios.

Effective use of neural representations for nearest neighbor retrieval.

Abstract

Duplicate question detection (DQD) is important to increase efficiency of community and automatic question answering systems. Unfortunately, gathering supervised data in a domain is time-consuming and expensive, and our ability to leverage annotations across domains is minimal. In this work, we leverage neural representations and study nearest neighbors for cross-domain generalization in DQD. We first encode question pairs of the source and target domain in a rich representation space and then using a k-nearest neighbour retrieval-based method, we aggregate the neighbors' labels and distances to rank pairs. We observe robust performance of this method in different cross-domain scenarios of StackExchange, Spring and Quora datasets, outperforming cross-entropy classification in multiple cases.