This paper explores the use of probabilistic and conventional qualitative spatial reasoning (QSR) in the context of geospatial question answering (GeoQA) systems. The paper presents a thorough empirical investigation of the performance of a probabilistic and a conventional qualitative spatial reasoner, across a range increasingly sophisticated scenarios with real data and synthetically generated questions. The results indicate the potential of probabilistic QSR to provide more detailed information about spatial configurations than conventional QSR; but at the cost of less frequent errors in estimating the relative likelihood of different reasoning conclusions. Errors in probabilistic reasoning also tend to be systematically associated with lower probability conclusions. The results have implications for reliable and flexible automated spatial reasoning systems, especially where neither conventional geographic information retrieval (GIR) techniques nor large language models (LLMs) are able to provide a satisfactory solution to GeoQA problems.

During navigation, people may deviate from an already planned path. Possible reasons include traffic congestion, difficulty in following instructions, or personal preferences, as evidenced by studies in human spatial cognition and transportation. Most path planning methods provide users with a single path, for example the shortest one, while a few give the user a set of paths to choose from. In the latter case, once a path from the given set is chosen, the rest are discarded. If the user were to deviate from the planned path, alternatives may not exist or may cause a significant delay to the user. Services such as Google Maps continuously monitor possible alternatives, but they may offer a path with no alternatives or few and lengthy ones. Taking a proactive approach to finding alternatives, we introduce themost flexible path– a path that maximises the number of alternatives the user has along a given path. This way, users may choose to take a different path midway and be more likely to have alternative paths available to them. Straightforward approaches to obtaining such a path have a prohibitive computational complexity. We introduce a set of algorithms that yield exact and approximate solutions to this problem. We then showcase the most flexible path’s time-saving reroutes in a traffic congestion scenario.

Probabilistic logics combine the ability to reason about complex scenes, with a rigorous approach to uncertainty. This paper explores the construction of probabilistic spatial logics through the combination of established qualitative spatial calculi together with Markov logic networks (MLNs). Qualitative spatial calculi provide the basis for automated representation and reasoning with complex spatial scenes; MLNs provide a rigorous basis for handling uncertainty and driving probabilistic inference. Our approach focuses specifically on the combination of an uncertain knowledge base with a certain spatial reasoning rule-base. The experiments explore how uncertain knowledge propagates through certain qualitative spatial inferences, using the specific example of reasoning with cardinal directions. The results provide a template for probabilistic qualitative spatial reasoning more generally, with applications to a wide range of common scenarios for situational awareness and automated reasoning under uncertainty.

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.

Haonan Li, Martin Tomko, Maria Vasardani, Timothy Baldwin. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Online navigation systems assume a person can follow a given route from origin to destination. Nonetheless, spatial cognition studies show that wayfinding is a complex, highly adaptive process and that route planning is incremental rather than prescriptive. Indeed, people may deviate from their originally chosen route for a number of reasons including navigation errors, especially when the environment is unfamiliar. Even in familiar places, certainty in wayfinding is highly unlikely to be completely achieved. Consequently, when people make a wrong turn or miss an exit, even the best reroute to the destination may add several minutes to the originally planned travel time. This work formally defines the novel problem of finding a path such that, when navigation errors occur, recovering is not as costly. We call this approach the most recoverable path. This subtle change in the route planning problem – i.e., considering error recovery costs – invalidates using dynamic programming, such as in shortest path algorithm solutions. We therefore introduce a novel, readily applicable, fast heuristic to this NP-hard problem. The benefits of the most recoverable path are manifold: long detours are avoided, actual travel time is reduced, and it is comparable to its shortest counterpart in terms of length.

Existing metonymy resolution approaches rely on features extracted from external resources like dictionaries and hand-crafted lexical resources.In this paper, we propose an end-to-end word-level classification approach based only on BERT, without dependencies on taggers, parsers, curated dictionaries of place names, or other external resources.We show that our approach achieves the state-of-the-art on 5 datasets, surpassing conventional BERT models and benchmarks by a large margin.We also show that our approach generalises well to unseen data.

Mobile phone ubiquity has allowed the implementation of a number of emergency-related evacuation aids. Yet, these applications still face a number of challenges in human-mobile interaction, namely: (1) lack of widely accepted mobile usability guidelines, (2) people’s limited cognitive capacity when using mobile phones under stress, and (3) difficulty recreating emergency scenarios as experiments for usability testing. This study is intended as an initial view into smartphone usability under emergency evacuations by compiling a list of experimental observations and setting the ground for future research in cognitively-informed spatial algorithms and app design.

The higher social attention and negotiation among road users while crossing the road is as much a challenge for self-driving cars as it is for pedestrians. Self-driving cars in their current state are not able to understand cues from other road users and are rather reactive to pedestrian behavior, which may result overall in a slower traffic flow. In this paper, a vehicle-pedestrian negotiation model is proposed describing the processing and exchange of negotiation cues from both parties in order to speed up the traffic flow. The motion strategy for the vehicle approaching the pedestrian is formulated in order to negotiate its best chance to pass first, a process that closely mimics the common scenarios of everyday negotiation on roads. Simulation results show an improvement in the overall travel time of the vehicles as compared with the current best practice behavior (always stop) of autonomous vehicles. The cost-benefit analysis of negotiation among both parties is also discussed in this paper.

Place descriptions are used in everyday communication as a common way to convey spatial information. Processing the information from place descriptions poses multiple significant challenges because these descriptions are written in natural language. In particular, corpora of place descriptions provide a plethora of human spatial knowledge beyond geographical information system, even if these descriptions refer to the same places in various ways. This article focuses on resolving ambiguous or synonymous place names from place descriptions by exploring the given relationships with other spatial features. It matches place names from multiple descriptions by developing a novel labelled graph matching process that relies solely on the comparison of string, linguistic and spatial similarities between identified places. This process uses unstructured place descriptions as an input, and produces a composite place graph with qualitative spatial relations from the descriptions. The performance of this novel process exceeds current toponym resolution by coping with non-gazetteered places.

In this paper, we review the current literature on geographic information retrieval based on place names. We focus on the positional uncertainties and the extent of vagueness frequently associated with place names in linguistic place descriptions and on the differences between common users’ perception and the way the geographic information services interpret place names. We argue that, despite some notable efforts from the scientific community, geographic information services still cannot unambiguously recognize and sufficiently perform spatial reasoning with place names used in linguistic expressions. We focus on three interrelated research areas: (1) the use of place names in gazetteers, (2) the use of formal models to reason with spatial relations and with the spatial extent of place names in linguistic place descriptions, and (3) Web-harvesting and crowd-sourcing techniques for identifying place names and their spatial extension from public and volunteer sources, such as social networks and photo-sharing sites. We identify some opportunities for synthesizing existing approaches that would expedite the process of intelligent communication about place names between services and users. We discuss the shortcomings of the current state of affairs in locating place names from place descriptions and identify new areas of importance for future research.