Geo-Encoder: A Chunk-Argument Bi-Encoder Framework for Chinese Geographic Re-Ranking

TL;DR

The paper introduces Geo-Encoder, a novel bi-encoder framework that enhances Chinese geographic re-ranking by effectively integrating geographical semantics, leading to significant improvements over existing methods.

Contribution

Geo-Encoder employs chunking of geographical spans and multi-task learning with asynchronous updates to better capture Chinese geographical semantics in re-ranking tasks.

Findings

Achieves a 6.22% increase in Hit@1 score on GeoTES dataset.

Significantly outperforms state-of-the-art baselines.

Effective integration of geographical semantics improves re-ranking accuracy.

Abstract

Chinese geographic re-ranking task aims to find the most relevant addresses among retrieved candidates, which is crucial for location-related services such as navigation maps. Unlike the general sentences, geographic contexts are closely intertwined with geographical concepts, from general spans (e.g., province) to specific spans (e.g., road). Given this feature, we propose an innovative framework, namely Geo-Encoder, to more effectively integrate Chinese geographical semantics into re-ranking pipelines. Our methodology begins by employing off-the-shelf tools to associate text with geographical spans, treating them as chunking units. Then, we present a multi-task learning module to simultaneously acquire an effective attention matrix that determines chunk contributions to extra semantic representations. Furthermore, we put forth an asynchronous update mechanism for the proposed addition task, aiming to guide the model capable of effectively focusing on specific chunks. Experiments on two distinct Chinese geographic re-ranking datasets, show that the Geo-Encoder achieves significant improvements when compared to state-of-the-art baselines. Notably, it leads to a substantial improvement in the Hit@1 score of MGEO-BERT, increasing it by 6.22% from 62.76 to 68.98 on the GeoTES dataset.