ERNIE-GeoL: A Geography-and-Language Pre-trained Model and its Applications in Baidu Maps

TL;DR

This paper introduces ERNIE-GeoL, a specialized geography-and-language pre-trained model that enhances geo-related tasks at Baidu Maps by integrating geographic knowledge into the pre-training process, leading to improved performance.

Contribution

The paper presents ERNIE-GeoL, a novel pre-trained model that incorporates geographic knowledge for better geo-related task performance, addressing limitations of generic PTMs.

Findings

ERNIE-GeoL outperforms generic PTMs on large-scale geo datasets.

Deployment of ERNIE-GeoL improved downstream task accuracy at Baidu Maps.

The model demonstrates effective learning of geography-language representations.

Abstract

Pre-trained models (PTMs) have become a fundamental backbone for downstream tasks in natural language processing and computer vision. Despite initial gains that were obtained by applying generic PTMs to geo-related tasks at Baidu Maps, a clear performance plateau over time was observed. One of the main reasons for this plateau is the lack of readily available geographic knowledge in generic PTMs. To address this problem, in this paper, we present ERNIE-GeoL, which is a geography-and-language pre-trained model designed and developed for improving the geo-related tasks at Baidu Maps. ERNIE-GeoL is elaborately designed to learn a universal representation of geography-language by pre-training on large-scale data generated from a heterogeneous graph that contains abundant geographic knowledge. Extensive quantitative and qualitative experiments conducted on large-scale real-world datasets demonstrate the superiority and effectiveness of ERNIE-GeoL. ERNIE-GeoL has already been deployed in production at Baidu Maps since April 2021, which significantly benefits the performance of various downstream tasks. This demonstrates that ERNIE-GeoL can serve as a fundamental backbone for a wide range of geo-related tasks.