Bimodal data, such as image-text pairs, has become increasingly prevalent in the digital era. The Hybrid Vector Query (HVQ) is an effective approach for querying data and recently garnered considerable attention from researchers. It calculates similarity scores objects represented by two vectors using a weighted sum of each individual vector's similarity, with query-specific parameter α to determine weight. Existing methods HVQ typically construct Approximate Nearest Neighbors Search (ANNS)...

Bimodal data, such as image-text pairs, has become increasingly prevalent in the digital era. The Hybrid Vector Query (HVQ) is an effective approach for querying data and recently garnered considerable attention from researchers. It calculates similarity scores objects represented by two vectors using a weighted sum of each individual vector's similarity, with query-specific parameter $\alpha$ to determine weight. Existing methods HVQ typically construct Approximate Nearest Neighbors Search...

Large Language Models (LLMs) are poised to play an increasingly important role in our lives, providing assistance across a wide array of tasks. In the geospatial domain, LLMs have demonstrated ability answer generic questions, such as identifying country's capital; nonetheless, their utility is hindered when it comes answering fine-grained questions about specific places, grocery stores or restaurants, which constitute essential aspects people's everyday lives. This mainly because places...

Pre-trained Foundation Models (PFMs) have ushered in a paradigm-shift AI, due to their ability learn general-purpose representations that can be readily employed downstream tasks. While PFMs been successfully adopted various fields such as NLP and Computer Vision, capacity handling geospatial data remains limited. This attributed the intrinsic heterogeneity of data, which encompasses different types, including points, segments regions, well multiple information modalities. The proliferation...

A geospatial Knowledge Graph (KG) is a heterogeneous information network, capable of representing relationships between spatial entities in machine-interpretable format, and has tremendous applications logistics social networks. Existing efforts to build KG, have mainly used sparse relationships, e.g., district located inside city, which provide only marginal benefits compared traditional database. In spite the substantial advances tasks link prediction knowledge graph completion,...

A geospatial database is today at the core of an ever increasing number services. Building and maintaining it remains challenging due to need merge information from multiple providers. Entity Resolution (ER) consists finding entity mentions different sources that refer same real world entity. In ER, entities are often represented using schemes subject incomplete inaccurate location, making ER deduplication daunting tasks. While tremendous advances have been made in traditional resolution...

Pre-trained Foundation Models (PFMs) have ushered in a paradigm-shift Artificial Intelligence, due to their ability learn general-purpose representations that can be readily employed wide range of downstream tasks. While PFMs been successfully adopted various fields such as Natural Language Processing and Computer Vision, capacity handling geospatial data answering urban questions remains limited. This attributed the intrinsic heterogeneity data, which encompasses different types, including...