Renyao Chen

Geographic entity representation learning (GERL) is an emerging method that represents natural, administrative divisions, road networks, and points of interest (POIs) in a low-dimensional continuous vector space. GERL provides a fundamental approach for geographic entities to underpin a variety of intelligent applications by learning their representation vectors to capture the semantics and interactions of the entities. Previous GERL methods mainly focus on the representation learning of the geographic entities that are seen at the time of training, which struggle to accurately generate representation vectors for the growing number of unseen geographic entities that were not involved in model training. To address this issue, this article proposes spatial meta-learning-based representation learning (SMRL), which integrates spatial subgraphs and meta-learning to improve the representation vectors of unseen geographic entities. Specifically, SMRL first designs a spatial-aware subgraph sampling module based on attributes and relationships of geographic entities to divide entities into spatial subgraphs. It develops a local-level representation module to learn entity features at the subgraph level. Finally, SMRL proposes a meta-learning-driven representation strategy that integrates meta-learning to learn the representation of unseen geographic entities. Extensive experiments show that the proposed SMRL method outperforms baselines with both higher accuracy and higher computational efficiency. This study provides new explorations for the representation of unseen geographic entities and offers methodological References for the various geographic applications.

The Geographic Knowledge Graph (GeoKG) serves as an effective method for organizing geographic knowledge, playing a crucial role in facilitating semantic interoperability across heterogeneous data sources. However, existing GeoKGs are limited by a lack of hierarchical modeling and insufficient coverage of geographic knowledge (e.g., limited entity types, inadequate attributes, and insufficient spatial relationships), which hinders their effective use and representation of semantic content. This paper presents HGeoKG, a hierarchical geographic knowledge graph that comprehensively models hierarchical structures, attributes, and spatial relationships of multi-type geographic entities. Based on the concept and construction methods of HGeoKG, this paper developed a dataset named HGeoKG-MHT-670K. Statistical analysis reveals significant regional heterogeneity and long-tail distribution patterns in HGeoKG-MHT-670K. Furthermore, extensive geographic knowledge reasoning experiments on HGeoKG-MHT-670K show that most knowledge graph embedding (KGE) models fail to achieve satisfactory performance. This suggests the need to accommodate spatial heterogeneity across different regions and improve the embedding quality of long-tail geographic entities. HGeoKG serves as both a reference for GeoKG construction and a benchmark for geographic knowledge reasoning, driving the development of geographical artificial intelligence (GeoAI).

Geographic knowledge graph representation learning embeds entities and relationships in geographic knowledge graphs into a low-dimensional continuous vector space, which serves as a basic method that bridges geographic knowledge graphs and geographic applications. Previous geographic knowledge graph representation methods primarily learn the vectors of entities and their relationships from their spatial attributes and relationships, which ignores various semantics of entities, resulting in poor embeddings on geographic knowledge graphs. This study proposes a two-stage multimodal geographic knowledge graph representation (MGKGR) model that integrates multiple kinds of semantics to improve the embedding learning of geographic knowledge graph representation. Specifically, in the first stage, a spatial feature fusion method for modality enhancement is proposed to combine the structural features of geographic knowledge graphs with two modal semantic features. In the second stage, a multi-level modality feature fusion method is proposed to integrate heterogeneous features from different modalities. By fusing the semantics of text and images, the performance of geographic knowledge graph representation is improved, providing accurate representations for downstream geographic intelligence tasks. Extensive experiments on two datasets show that the proposed MGKGR model outperforms the baselines. Moreover, the results demonstrate that integrating textual and image data into geographic knowledge graphs can effectively enhance the model’s performance.

Region statistical data is widely utilized as fundamental spatial information for urban status observation, urban planning, and the formulation of urban policies. In practice, the collected region statistical data is usually incomplete due to insufficient sensors, recording errors or privacy restrictions. Transfer learning-based methods make great advancement on completing missing region statistical data in the data-sparse city by training model on the data-rich city. However, these methods employ all available attributes of regions in their dataset to learn transferred knowledge, even if the values of some attributes in the source and target cities differ significantly. To address this issue, this paper proposes a double graph neural network model to improve completion of the region statistical data on insufficient data cities. The model first introduces spatial distribution of attributes to identify transferable attributes, then utilizes hierarchical graph structure to characterize the spatial relationships of regions, finally designs a double graph network to transfer knowledge for completing the missing values of insufficient data cities. Experiments on two datasets show that the proposed model outperforms the baseline models and is robust. This study presents a promising method for spatial data completion, and provides a method reference for spatial applications with insufficient data.

Geographical entity representation learning (GERL) aims to embed geographical entities into a low-dimensional vector space, which provides a generalized approach for utilizing geographical entities to serve various geographical intelligence applications. In practice, the spatial distribution of geographical entities is highly unbalanced; thus, it is challenging to embed them accurately. Previous GERL models treated all geographical entities uniformly, resulting in insufficient entity representations. To address this issue, this article proposes an anchor-enhanced GERL (AE-GERL) model, which utilizes the key informative entities as anchors to improve the representations of geographical entities. Specifically, AE-GERL develops an anchor selection algorithm to identify anchors from large-scale geographical entities based on their spatial distribution and entity types. To utilize anchors to guide geographical entities, AE-GERL constructs an anchor-enhanced graph to establish explicit connections between anchors and nonanchor entities. Finally, a graph neural network (GNN) based anchor to nonanchor node learning model is designed to impute missing information of nonanchor entities. Extensive experiments are conducted on four datasets, and the experimental results demonstrate that AE-GERL outperforms the baseline models in both sparse and dense scenarios. This study provides a methodological reference for embedding geographical entities in various geographical applications and also provides an effective approach to improve the performance of message-passing-based GNN models.

Human activities embedded in crowdsourced data, such as social media trajectory, represent individual daily styles and patterns, which are valuable in many applications. However, the accurate identification of human activity types (HATs) from social media is challenging, possibly because interactions between posts and users at different time are overlooked. To fill this gap, we propose a novel model that introduces the interactions hidden in social media and synthesizes Graph Convolutional Network (GCN) for identifying HAT. The model first characterizes interactions among words, posts, dates, and users, and then derives a Time Gated Human Activity Graph Convolutional Network (TG-HAGCN) to predict the HATs of social media trajectory. To examine the proposed model performance, we built a new dataset including interactions between post content, post time, and users from the open Yelp dataset. Experimental results show that exploiting interactions hidden in social media to recognize HATs achieves state-of-the-art performance with high accuracy. The study indicates that interactions among social media promotes ability of machine learning on social media data mining and intelligent applications, and offers a reference solution for how to fuse multi-type heterogeneous data in social media.

Word embedding is an important reference for natural language processing tasks, which can generate distribution presentations of words based on many text data. Recent evidence demonstrates that introducing sememe knowledge is a promising strategy to improve the performance of word embedding. However, previous works ignored the structure information of sememe knowledges. To fill the gap, this study implicitly synthesized the structural feature of sememes into word embedding models based on an attention mechanism. Specifically, we propose a novel double attention word-based embedding (DAWE) model that encodes the characteristics of sememes into words by a “double attention” strategy. DAWE is integrated with two specific word training models through context-aware semantic matching techniques. The experimental results show that, in word similarity task and word analogy reasoning task, the performance of word embedding can be effectively improved by synthesizing the structural information of sememe knowledge. The case study also verifies the power of DAWE model in word sense disambiguation task. Furthermore, the DAWE model is a general framework for encoding sememes into words, which can be integrated into other existing word embedding models to provide more options for various natural language processing downstream tasks.