Qualitative Spatial Reasoning over Questions (Short Paper)
Конвейер с поддержкой искусственного интеллекта для динамической генерации достоверного контента о пищевых добавках в больших масштабах
2022-01-01
SCID: 54.1/3pwdz8fv
Discuss with AI
cardinal direction calculusgeoparsinggeospatial question answeringqualitative spatial question answeringspatial reasoning
Figures from the paper
Abstract (AI)
Although geospatial question answering systems have received increasing attention in recent years, existing prototype systems struggle to properly answer qualitative spatial questions. In this work, we propose a unique framework for answering qualitative spatial questions, which comprises three main components: a geoparser that takes the input questions and extracts place semantic information from text, a reasoning system which is embedded with a crisp reasoner, and finally, answer extraction, which refines the solution space and generates final answers. We present an experimental design to evaluate our framework for point-based cardinal direction calculus (CDC) relations by developing an automated approach for generating three types of synthetic qualitative spatial questions. The initial evaluations of generated answers in our system are promising because a high proportion of answers were labelled correct.
Key Findings
1
Initial experiments indicate that a high proportion of the generated answers are correctly labeled, suggesting promising framework performance.
2
The geoparser extracts semantic information about places from natural-language questions to support downstream qualitative spatial reasoning.
3
The paper proposes a three-component framework for qualitative spatial question answering: geoparsing, crisp spatial reasoning, and answer extraction.
4
The system is evaluated on point-based cardinal direction calculus relations using an automated generator for three types of synthetic spatial questions.
Research Object
Framework for answering qualitative geospatial (spatial) questions comprising geoparser, crisp reasoning system, and answer extraction
Research Subject
the framework’s ability to extract place semantics, reason over qualitative spatial relations, and generate correct answers to synthetic CDC questions
Publication Details
Publication Date
2022-01-01
Journal
Publisher
ISSN
Cited by
1
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai3
Cited by20
Transformers: State-of-the-Art Natural Language Processing2020
Graph neural networks: A review of methods and applications2020
Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting2021
A Survey on Evaluation of Large Language Models2024
Emerging Properties in Self-Supervised Vision Transformers2021
Are Transformers Effective for Time Series Forecasting?2023
Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers2021
UNETR: Transformers for 3D Medical Image Segmentation2022
Transformer-XL: Attentive Language Models beyond a Fixed-Length Context2019
HuggingFace's Transformers: State-of-the-art Natural Language Processing2019
CodeBERT: A Pre-Trained Model for Programming and Natural Languages2020
Point Transformer2021
Swin Transformer V2: Scaling Up Capacity and Resolution2022
Unifying Large Language Models and Knowledge Graphs: A Roadmap2024
ChatGPT: A comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope2023
Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet2021
Pre-Trained Image Processing Transformer2021
CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification2021
WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing2022
EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks2019