Do Sentence Transformers Learn Quasi-Geospatial Concepts from General Text?
This work addresses the problem of enhancing routing recommendation systems for hikers by leveraging pre-trained models, though it is incremental as it builds on existing sentence transformer methods.
The study investigated whether sentence transformers fine-tuned on general text can understand quasi-geospatial concepts like route types and difficulty from descriptions of hiking routes in Great Britain, finding they have some zero-shot capabilities for such tasks.
Sentence transformers are language models designed to perform semantic search. This study investigates the capacity of sentence transformers, fine-tuned on general question-answering datasets for asymmetric semantic search, to associate descriptions of human-generated routes across Great Britain with queries often used to describe hiking experiences. We find that sentence transformers have some zero-shot capabilities to understand quasi-geospatial concepts, such as route types and difficulty, suggesting their potential utility for routing recommendation systems.