Extreme multilabel classification (XMLC) problems occur in settings such as related product recommendation, large-scale document tagging, or ad prediction, and are characterized by a label space that can span millions of possible labels. There are two implicit tasks that the classifier performs: Evaluating each potential label for its expected worth, and then selecting the best candidates. For the latter task, only the relative order of scores matters, and this is what is captured by the standard evaluation procedure in the XMLC literature. However, in many practical applications, it is important to have a good estimate of the actual probability of a label being relevant, e.g., to decide whether to pay the fee to be allowed to display the corresponding ad. To judge whether an extreme classifier is indeed suited to this task, one can look, for example, to whether it returns calibrated probabilities, which has hitherto not been done in this field. Therefore, this paper aims to establish the current status quo of calibration in XMLC by providing a systematic evaluation, comprising nine models from four different model families across seven benchmark datasets. As naive application of Expected Calibration Error (ECE) leads to meaningless results in long-tailed XMC datasets, we instead introduce the notion of calibration@k (e.g., ECE@k), which focusses on the top-k probability mass, offering a more appropriate measure for evaluating probability calibration in XMLC scenarios. While we find that different models can exhibit widely varying reliability plots, we also show that post-training calibration via a computationally efficient isotonic regression method enhances model calibration without sacrificing prediction accuracy. Thus, the practitioner can choose the model family based on accuracy considerations, and leave calibration to isotonic regression.

Accurate and rapid localization can improve geographic information system (GIS) tasks to support disaster rescue and resource exploration in field operations. However, the existing localization methods suffer from high costs, low efficiency, and poor accuracy of long-distance targets due to environmental factors like terrain and landforms. Therefore, we propose a multilevel localization framework based on mobile phones for different field scenarios. First, we investigated a rapid localization method constrained by scene contextual features in the field with salient features. Second, we designed a single-point localization method that combines DEM data and mobile phones when high-precision DEM data are available for the region. Third, we studied a map-matching corrected joint localization with mobile phone images in the field lacking salient features and high-precision DEM data. Finally, we developed a prototype system for field localization and selected a forest scene for experimental analysis. The results showed that the proposed mobile phone-based localization framework supports long-distance localization in the field. The localization accuracy in three different field scenarios is within 100 meters, and the localization efficiency reaches a minute level, which can effectively support the convenient, rapid, high-precision, and long-distance localization tasks in field scenes.

Efficient geovisualization is beneficial for understanding geospatial phenomena, an important research direction for GISers and Cartographers. However, the current research on geovisualization overemphasizes the visual effects while neglecting the prominent representation of crucial information and failing to consider the user’s cognitive workload of information processing. Following the laws of visual perception of the human eyes, this article proposes a visual attention-guided augmented representation approach of geographic scenes that involves area of interest computation, background simplification, and compound graphic variables. Finally, we select bridge stress visualization as a case study for experimental analysis. The experimental results of eye-tracking show that augmented representation could draw the participants’ attention to areas of interest in a short time, increasing their duration of fixations and the accuracy of completing given tasks. These findings suggest that our approach can enhance geographic scenes’ cognitive efficiency, offers a new idea for the theoretical studies of geovisualization, and holds promising potential for broader application in various geographical phenomena visualization.

An important step in implementing geospatial digital twins is to enhance the expressiveness of virtual geographical scenes for the physical world. However, the existing virtual geographical scenes cannot quickly express the dynamically changing geographic environment for remote users due to the inefficient handling of modeling processes, user perception, and remote sharing. The research analysed the concept and characteristics of geospatial digital twins, and constructed the virtual geographical scene ontology, based on which we developed geographical spatiotemporal semantic rules and designed a dynamic annotation algorithm to enhance the representation of virtual geographical scenes. Finally, we investigated a real-time transmission method of panoramic video based on 5 G and used immersive virtual reality (IVR) to realize the user experience of remote immersion in geographical scenes. We selected a specific geographic environment containing multiple typical geographic entities to develop three prototype systems for experimental analyses. The results showed that the proposed method enabled users to view the virtual geographical scene on a VR device. The average latency for this process was 14.72 seconds. Compared with the virtual geographical scenes constructed by traditional methods, the experiments showed the proposed method advantageous in comprehensiveness, timeliness, and photorealism and abilities to enhance the user’s geographical scene perception.

The detail and representation of a spatial layout varies with scale. This affects an individual’s learning effectiveness and understanding, in turn directly influencing their behavior in a fire evacuation. However, the impact of layout learning methods with different spatial scales on fire evacuation behavior, and the relationship between spatial cognition and evacuation effects, remains unclear. We conducted spatial layout learning across three scales with 81 participants and simulated a fire evacuation scenario in a mobile virtual reality for groups. We collected evacuation decision-making and user experience questionnaires as supplementary data. The results demonstrate that small-scale learning objects are the easiest for participants to understand in terms of spatial layout and relationships, but their performance in fire evacuation is poor. Large-scale learning objects significantly improve participants’ evacuation efficiency. Spatial layout learning plays a crucial role in fire evacuation outcomes, but traditional spatial knowledge acquisition measurement methods cannot predict fire evacuation performance. This study sheds light on how spatial cognition influences fire evacuation behavior and provides a more reliable fire evacuation simulation method based on mobile virtual reality (MVR).

As an important application of virtual geographic environments (VGEs), virtual disaster scenes are essential in enhancing the public’s risk awareness. However, existing virtual disaster scene visualization methods lack expert guidance and fail to meet the public’s requirements, resulting in an ineffective public understanding. Therefore, this paper proposes a knowledge-guided disaster scene 3D visualization framework. First, the public’s demand for disaster scene visualization is analyzed, and a geographic knowledge graph of disaster scenes is constructed. Second, through the guidance of the knowledge graph, the virtual disaster scenes are fusion modeled and suitability represented. Third, a diverse organization and adaptive scheduling method of disaster scene data for multi-computing devices is established. Finally, we developed a prototype system for disaster scene visualization, selected a typical disaster, and conducted cognitive experiments with eye-tracking technology. The results show that the proposed method can effectively support the adaptive visualization of virtual disaster scenes for four computing devices and maintain an efficient frame rate. In addition, compared with other disaster scene visualization methods, our framework incorporates semantic knowledge of scene, user, demand, and space. It can effectively convey disaster information and help the public cognize disaster risks and has significant advantages in modeling standardization, personalization, and adaptability.