Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Location Aware Modular Biencoder for Tourism Question Answering

Haonan Li, Martin Tomko, Timothy Baldwin

TL;DR

This paper tackles real-world tourism question answering by addressing the need for joint spatial and textual reasoning over vast POI pools. It introduces Lamb, a location-aware modular bi-encoder that encodes questions and POIs separately and fuses textual and geospatial signals for efficient dense retrieval. A two-phase training regime with varied negative sampling, coupled with SelSum-based POI pre-processing, yields state-of-the-art results and scalable global evaluation. The approach demonstrates strong accuracy and practical efficiency, making it suitable for deployment in large-scale tourism applications while highlighting dataset limitations and avenues for future geo-aware enhancements.

Abstract

Answering real-world tourism questions that seek Point-of-Interest (POI) recommendations is challenging, as it requires both spatial and non-spatial reasoning, over a large candidate pool. The traditional method of encoding each pair of question and POI becomes inefficient when the number of candidates increases, making it infeasible for real-world applications. To overcome this, we propose treating the QA task as a dense vector retrieval problem, where we encode questions and POIs separately and retrieve the most relevant POIs for a question by utilizing embedding space similarity. We use pretrained language models (PLMs) to encode textual information, and train a location encoder to capture spatial information of POIs. Experiments on a real-world tourism QA dataset demonstrate that our approach is effective, efficient, and outperforms previous methods across all metrics. Enabled by the dense retrieval architecture, we further build a global evaluation baseline, expanding the search space by 20 times compared to previous work. We also explore several factors that impact on the model's performance through follow-up experiments. Our code and model are publicly available at https://github.com/haonan-li/LAMB.

Location Aware Modular Biencoder for Tourism Question Answering

TL;DR

This paper tackles real-world tourism question answering by addressing the need for joint spatial and textual reasoning over vast POI pools. It introduces Lamb, a location-aware modular bi-encoder that encodes questions and POIs separately and fuses textual and geospatial signals for efficient dense retrieval. A two-phase training regime with varied negative sampling, coupled with SelSum-based POI pre-processing, yields state-of-the-art results and scalable global evaluation. The approach demonstrates strong accuracy and practical efficiency, making it suitable for deployment in large-scale tourism applications while highlighting dataset limitations and avenues for future geo-aware enhancements.

Abstract

Answering real-world tourism questions that seek Point-of-Interest (POI) recommendations is challenging, as it requires both spatial and non-spatial reasoning, over a large candidate pool. The traditional method of encoding each pair of question and POI becomes inefficient when the number of candidates increases, making it infeasible for real-world applications. To overcome this, we propose treating the QA task as a dense vector retrieval problem, where we encode questions and POIs separately and retrieve the most relevant POIs for a question by utilizing embedding space similarity. We use pretrained language models (PLMs) to encode textual information, and train a location encoder to capture spatial information of POIs. Experiments on a real-world tourism QA dataset demonstrate that our approach is effective, efficient, and outperforms previous methods across all metrics. Enabled by the dense retrieval architecture, we further build a global evaluation baseline, expanding the search space by 20 times compared to previous work. We also explore several factors that impact on the model's performance through follow-up experiments. Our code and model are publicly available at https://github.com/haonan-li/LAMB.
Paper Structure (34 sections, 6 equations, 4 figures, 9 tables)

This paper contains 34 sections, 6 equations, 4 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Task Formulation
  4. POI Pre-processing
  5. Location Aware Modular Bi-encoder
  6. Textual Module
  7. Location Module
  8. Question and POI Encoders
  9. Training and Inference
  10. Experimental Setup
  11. Dataset
  12. Evaluation Metrics
  13. Baselines
  14. Lamb Implementation Details
  15. Results and Analysis
  16. ...and 19 more sections

Figures (4)

  • Figure 1: An example of real-world POI recommendation question from the TourismQA dataset contractor2019large. Colored text represents constraints relevant to recommending POIs.
  • Figure 2: Proposed approach. The reviews of POIs are first selected and summarized by SelSum (bottom right part). A location module is separately pre-trained under the supervision of geocoordinate-based distances (top right part). The left part is the main Lamb model. The cyan- and salmon-coloured parts are the POI encoder and question encoder, respectively. The orange part is the index of POI embeddings used for inference.
  • Figure 3: Example of SelSum output.
  • Figure 4: Top-$k$ accuracy with varying numbers of training epochs.