Nearest-neighbor Method Research Articles

Efficient Feature Matching for Large-scale Images based on Cascade Hash and Local Geometric Constraint

Abstract. Feature matching plays a crucial role in 3D reconstruction to provide correspondences between overlapped images. The accuracy and efficiency of feature matching significantly impact the performance of 3D reconstruction. The widely used framework with the exhaustive nearest neighbor searching (NNS) between descriptors and RANSAC-based geometric estimation is, however, low-efficient and unreliable for large-scale UAV images. Inspired by indexing-based NNS, this paper implements an efficient feature matching method for large-scale images based on Cascade Hashing and local geometric constraints. Our proposed method improves upon traditional feature matching approaches by introducing a combination of image retrieval, data scheduling, and GPU-accelerated Cascade Hashing. Besides, it utilizes a local geometric constraint to filter matching results within a matching framework. On the one hand, the GPU-accelerated Cascade Hashing technique generates compact and discriminative hash codes based on image features, facilitating the rapid completion of the initial matching process, and significantly reducing the search space and time complexity. On the other hand, after the initial matching is completed, the method employs a local geometric constraint to filter the initial matching results, enhancing the accuracy of the matching results. This forms a three-tier framework based on data scheduling, GPU-accelerated Cascade Hashing, and local geometric constraints. We conducted experiments using two sets of large-scale UAV image data, comparing our method with SIFTGPU to evaluate its performance in initial matching, outlier rejection, and 3D reconstruction. The results demonstrate that our method achieves a feature matching speed 2.0 times that of SIFTGPU while maintaining matching accuracy and producing comparable reconstruction results. This suggests that our method holds promise for efficiently addressing large-scale image matching.

Large-scale DSM registration via motion averaging

Abstract. Generating wide-area digital surface models (DSMs) requires registering a large number of individual, and partially overlapped DSMs. This presents a challenging problem for a typical registration algorithm, since when a large number of observations from these multiple DSMs are considered, it may easily cause memory overflow. Sequential registration algorithms, although can significantly reduce the computation, are especially vulnerable for small overlapped pairs, leading to a large error accumulation. In this work, we propose a novel solution that builds the DSM registration task as a motion averaging problem: pair-wise DSMs are registered to build a scene graph, with edges representing relative poses between DSMs. Specifically, based on the grid structure of the large DSM, the pair-wise registration is performed using a novel nearest neighbor search method. We show that the scene graph can be optimized via an extremely fast motion average algorithm with O(N) complexity (N refers to the number of images). Evaluation of high-resolution satellite-derived DSM demonstrates significant improvement in computation and accuracy.

Establishment and Verification of an Artificial Intelligence Prediction Model for Children With Sepsis.

Early identification of high-risk groups of children with sepsis is beneficial to reduce sepsis mortality. This article used artificial intelligence (AI) technology to predict the risk of death effectively and quickly in children with sepsis in the pediatric intensive care unit (PICU). This retrospective observational study was conducted in the PICUs of the First Affiliated Hospital of Sun Yat-sen University from December 2016 to June 2019 and Shenzhen Children's Hospital from January 2019 to July 2020. The children were divided into a death group and a survival group. Different machine language (ML) models were used to predict the risk of death in children with sepsis. A total of 671 children with sepsis were enrolled. The accuracy (ACC) of the artificial neural network model was better than that of support vector machine, logical regression analysis, Bayesian, K nearest neighbor method and decision tree models, with a training set ACC of 0.99 and a test set ACC of 0.96. The AI model can be used to predict the risk of death due to sepsis in children in the PICU, and the artificial neural network model is better than other AI models in predicting mortality risk.

Abstract PO2-09-06: Interaction of racial disparities on outcomes and toxicities associated with treatment of HER2+Breast Cancer- a TrinetX Database study

Abstract Objective: Investigate the interaction of racial disparities on outcomes and toxicities associated with treatment of HER2+BrCa (Breast Cancer)- a TrinetX Database study Background: HER2+BrCa is an aggressive subtype accounting for about 1/3rd of all BrCa, for which Trastuzumab based therapy remains the mainstay of treatment. While we are starting to understand the breadth of racial disparities in BrCa , the knowledge about outcomes and toxicities in relation to treatment of HER2+ BrCa is still limited. Design/Methods: In this propensity score-matched cohort study we used the TriNetX Research Network, a multi-health care organization de-identified electronic health record (EHR) database, to compare the 1-, 3-, and 5-year mortality of HER2+BrCa non-Hispanic African American (NHAA) women to a corresponding non-Hispanic White cohort (NHW). Qualification into the two race based HER2+BrCa cohorts required the presence of a C50 ICD-10-CM diagnosis code and at least one proxy for HER2 positivity such as Trastuzumab administration (index event). Cohorts were matched for age, BMI, comorbidities (HTN, DM ), and lab values using 1:1 matching with a greedy nearest neighbor search. Toxicity outcomes of interest such as cardiomyopathy, neuropathy, diarrhea, rashes, leg swelling/edema, and fatigue as well as the frequency of emergency room visits were also compared between the cohorts. The associations of observed outcome frequencies in the two cohorts were tested for significance using the chi-square test. The odds ratios and p-values, corrected for multiple hypothesis testing using the Benjamini-Hochberg procedure, are reported as an effect size and significance estimation. Results: 15983 patients met the inclusion criteria for HER2+BrCa (13611-NHW and 2372-NHAA). The mean age at the index event was 55.6 ± 12.6 years for NHAA and 57.2 ± 13.3 years for NHW. After matching the two sub-cohorts, 2303 patients remained in each. NHAA women were found to have significantly increased odds of neuropathy and cardiomyopathy for 1, 3, and 5-year time intervals after the HER2+BrCa index date, as compared to the reference NHW women (abbreviated results given in Table 1). In addition, the odds of an emergency room visit (for any reason) was found to be increased by as much as 79% in NHAA women as compared to the NHW women. The odds of experiencing fatigue/malaise, diarrhea/nausea, and rash, were reduced for NHAA women across all time intervals investigated. Odds ratios for mortality and leg swelling did not meet the significance threshold (p-value < 0.05). Conclusion: This study, which examined racial disparities in outcomes and toxicities associated with HER2+BrCa, revealed significant findings. NHAA women with HER2+BrCa treated with Trastuzumab based therapy demonstrated increased odds of experiencing neuropathy and cardiomyopathy compared to their NHW counterparts. Additionally, NHAA women had a higher likelihood of emergency room visits for any reason. Conversely, NHAA women exhibited decreased odds of fatigue/malaise, diarrhea/nausea, and rash. These findings underscore the importance of considering the impact of race on HER2+BrCa outcomes and treatment toxicities and suggest the need for further research and targeted interventions to address disparities and thus enhance treatment outcomes in diverse populations. Table 1: Representative significant Odds ratios and corrected p-values for outcomes comparing the NHAA and NHW cohorts. Citation Format: Monali Vasekar, Justin Petucci, Avnish Katoch, Vasant Honavar. Interaction of racial disparities on outcomes and toxicities associated with treatment of HER2+Breast Cancer- a TrinetX Database study [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-09-06.

DET-LSH: A Locality-Sensitive Hashing Scheme with Dynamic Encoding Tree for Approximate Nearest Neighbor Search

Locality-sensitive hashing (LSH) is a well-known solution for approximate nearest neighbor (ANN) search in high-dimensional spaces due to its robust theoretical guarantee on query accuracy. Traditional LSH-based methods mainly focus on improving the efficiency and accuracy of the query phase by designing different query strategies, but pay little attention to improving the efficiency of the indexing phase. They typically fine-tune existing data-oriented partitioning trees to index data points and support their query strategies. However, their strategy to directly partition the multi-dimensional space is time-consuming, and performance degrades as the space dimensionality increases. In this paper, we design an encoding-based tree called Dynamic Encoding Tree (DE-Tree) to improve the indexing efficiency and support efficient range queries based on Euclidean distance. Based on DE-Tree, we propose a novel LSH scheme called DET-LSH. DET-LSH adopts a novel query strategy, which performs range queries in multiple independent index DE-Trees to reduce the probability of missing exact NN points, thereby improving the query accuracy. Our theoretical studies show that DET-LSH enjoys probabilistic guarantees on query accuracy. Extensive experiments on real-world datasets demonstrate the superiority of DET-LSH over the state-of-the-art LSH-based methods on both efficiency and accuracy. While achieving better query accuracy than competitors, DET-LSH achieves up to 6x speedup in indexing time and 2x speedup in query time over the state-of-the-art LSH-based methods.

Metabolomics Signature in Prediabetes and Diabetes: Insights From Tandem Mass Spectrometry Analysis.

This study investigates the metabolic differences between normal, prediabetic and diabetic patients with good and poor glycaemic control (GGC and PGC). In this study, 1102 individuals were included, and 50 metabolites were analysed using tandem mass spectrometry. The diabetes diagnosis and treatment standards of the American Diabetes Association (ADA) were used to classify patients. The nearest neighbour method was used to match controls and cases in each group on the basis of age, sex and BMI. Factor analysis was used to reduce the number of variables and find influential underlying factors. Finally, Pearson's correlation coefficient was used to check the correlation between both glucose and HbAc1 as independent factors with binary classes. Amino acids such as glycine, serine and proline, and acylcarnitines (AcylCs) such as C16 and C18 showed significant differences between the prediabetes and normal groups. Additionally, several metabolites, including C0, C5, C8 and C16, showed significant differences between the diabetes and normal groups. Moreover, the study found that several metabolites significantly differed between the GGC and PGC diabetes groups, such as C2, C6, C10, C16 and C18. The correlation analysis revealed that glucose and HbA1c levels significantly correlated with several metabolites, including glycine, serine and C16, in both the prediabetes and diabetes groups. Additionally, the correlation analysis showed that HbA1c significantly correlated with several metabolites, such as C2, C5 and C18, in the controlled and uncontrolled diabetes groups. These findings could help identify new biomarkers or underlying markers for the early detection and management of diabetes.

Endoscopic Cryoablation Versus Radical Nephroureterectomy for Upper Tract Urothelial Carcinoma

Background and objectiveCryoablation is a traditional antitumor therapy with good prospects for development. The efficacy of endoscopic management as a kidney-sparing surgery for high-risk upper tract urothelial carcinoma (UTUC) remains controversial. Our aim was to evaluate the impact of endoscopic cryoablation (ECA) versus radical nephroureterectomy (RNU) on survival outcomes, renal function, and complications. MethodsWe retrospectively analyzed data for 116 patients with newly diagnosed high-risk UTUC who underwent either ECA (n = 13) or RNU (n = 103) from March 25, 2019 to December 8, 2021. Propensity score matching (1:4) using the nearest neighbor method was performed before analysis. The primary outcome was overall survival (OS). Secondary outcomes included progression-free survival (PFS), intravesical recurrence-free survival (RFS), the change in renal function, and treatment-emergent adverse events (TEAEs). Key findings and limitationsAt median follow-up of 28.2 mo for the ECA group and 27.6 mo for the RNU group, 2-yr OS (82% vs 84%), PFS (73% vs 71%), and intravesical RFS (81% vs 83%) rates after matching did not significantly differ. A decline in renal function was observed after RNU, but not after ECA. Five (41.7%) patients in the ECA group reported six TEAEs, and 17 patients (35.4%) in the RNU group reported 20 TEAEs. Conclusions and clinical implicationsIn comparison to RNU, ECA for UTUC resulted in noninferior oncological outcomes and superior preservation of renal function. Patient summaryOur study suggests that a treatment called endoscopic cryoablation for high-risk cancer in the upper urinary tract can help in preserving kidney function, with similar survival outcomes to those after more extensive surgery. This option can be considered for selected patients with a strong preference for kidney preservation.

From Text to Recommendations: How Vector Databases are Revolutionizing Personalized Content Delivery

Abstract: Effective similarity search and retrieval are now possible thanks to vector databases, which have become a potent tool for organizing and searching high-dimensional data. This article provides a thorough examination of vector databases, their underlying theories, and their applications across a range of industries. In handling complex data types, we address the significance of vector representation and emphasize the benefits of vector databases over traditional databases [1]. The article explores the process of creating a vector database, highlighting the critical function of indexing strategies such as IVF (Inverted File Indexing) and HNSW (Hierarchical Navigable Small World) in guaranteeing the effectiveness and precision of searches [2]. In addition, we discuss the problems caused by the curse of dimensionality and offer solutions to lessen its effects on nearest neighbor searches [3]

Prediction of Myocardial Infarction Using a Combined Generative Adversarial Network Model and Feature-Enhanced Loss Function.

Accurate risk prediction for myocardial infarction (MI) is crucial for preventive strategies, given its significant impact on global mortality and morbidity. Here, we propose a novel deep-learning approach to enhance the prediction of incident MI cases by incorporating metabolomics alongside clinical risk factors. We utilized data from the KORA cohort, including the baseline S4 and follow-up F4 studies, consisting of 1454 participants without prior history of MI. The dataset comprised 19 clinical variables and 363 metabolites. Due to the imbalanced nature of the dataset (78 observed MI cases and 1376 non-MI individuals), we employed a generative adversarial network (GAN) model to generate new incident cases, augmenting the dataset and improving feature representation. To predict MI, we further utilized multi-layer perceptron (MLP) models in conjunction with the synthetic minority oversampling technique (SMOTE) and edited nearest neighbor (ENN) methods to address overfitting and underfitting issues, particularly when dealing with imbalanced datasets. To enhance prediction accuracy, we propose a novel GAN for feature-enhanced (GFE) loss function. The GFE loss function resulted in an approximate 2% improvement in prediction accuracy, yielding a final accuracy of 70%. Furthermore, we evaluated the contribution of each clinical variable and metabolite to the predictive model and identified the 10 most significant variables, including glucose tolerance, sex, and physical activity. This is the first study to construct a deep-learning approach for producing 7-year MI predictions using the newly proposed loss function. Our findings demonstrate the promising potential of our technique in identifying novel biomarkers for MI prediction.

Intercomparisons of Tracker v1.1 and four other ocean particle-tracking software packages in the Regional Ocean Modeling System

Abstract. Particle tracking is widely utilized to study transport features in a range of physical, chemical, and biological processes in oceanography. In this study, a new offline particle-tracking package, Tracker v1.1, is introduced, and its performance is evaluated in comparison to an online Eulerian dye, one online particle-tracking software package, and three offline particle-tracking software packages in a small, high-resolution model domain and a large coarser model domain. It was found that both particle and dye approaches give similar results across different model resolutions and domains when they were tracking the same water mass, as indicated by similar mean advection pathways and spatial distributions of dye and particles. The flexibility of offline particle tracking and its similarity against online dye and online particle tracking make it a useful tool to complement existing ocean circulation models. The new Tracker was shown to be a reliable particle-tracking package to complement the Regional Ocean Modeling System (ROMS) with the advantages of platform independence and speed improvements, especially in large model domains achieved by the nearest-neighbor search algorithm. Lastly, trade-offs of computational efficiency, modifiability, and ease of use that can influence the choice of which package to use are explored. The main value of the present study is that the different particle and dye tracking codes were all run on the same model output or within the model that generated the output. This allows some measure of intercomparison between the different tracking schemes, and we conclude that all choices that make each tracking package unique do not necessarily lead to very different results.

Correcting the Location Error of Persistent Scatterers in an Urban Area Based on Adaptive Building Contours Matching: A Case Study of Changsha

Persistent Scatterer InSAR (PS-InSAR) technology enables the monitoring of displacement in millimeters. However, without the use of external parameter correction, radar scatterers exhibit poor geopositioning precision in meters, limiting the correlation between observed deformation and the actual structure. The integration of PS-InSAR datasets and building databases is often overlooked in deformation research. This paper presents a novel strategy for matching between PS points and building contours based on spatial distribution characteristics. A convex hull is employed to simplify the building outline. Considering the influence of building height and incident angle on geometric distortion, an adaptive buffer zone is established. The PS points on a building are further identified through the nearest neighbor method. In this study, both ascending and descending TerraSAR-X orbit datasets acquired between 2016 and 2019 were utilized for PS-InSAR monitoring. The efficacy of the proposed method was evaluated by comparing the PS-InSAR results obtained from different orbits. Through a process of comparison and verification, it was demonstrated that the matching effect between PS points and building contours was significantly enhanced, resulting in an increase of 29.2% in the number of matching PS points. The results indicate that this novel strategy can be employed to associate PS points with building outlines without the need for complex calculations, thereby providing a robust foundation for subsequent building risk assessment.

Automatic inference of ICD-10 codes from German ophthalmologic physicians’ letters using natural language processing

Physicians’ letters are the optimal source of diagnoses for registries. However, most registries demand for diagnosis codes such as ICD-10. We herein describe an algorithm that infers ICD-10 codes from German ophthalmologic physicians’ letters. We assess the method in three German eye hospitals. Our algorithm is based on the nearest-neighbor method as well as on a large thesaurus for ICD-10 codes. This thesaurus was embedded into a Word2Vec space created from anonymized physicians’ reports of the first hospital. For evaluation, each of the three hospitals sent all diagnoses taken from 100 letters. The inferred ICD-10 codes were evaluated for correctness by the senders. A total of 3332 natural language terms had been sent in (812 hospital one, 1473 hospital two, 1047 hospital three). A total of 526 non-diagnoses were excluded upfront. 2806 ICD-10 codes were inferred (771 hospital one, 1226 hospital two, 809 hospital three). In the first hospital, 98% were fully correct and 99% correct at the level of the superordinate disease concept. The percentages in hospital two were 69% and 86%. The respective numbers for hospital three were 69% and 91%. Our simple method is capable of inferring ICD-10 codes for German natural language diagnoses, especially when the embedding space has been built with physicians’ letters from the same hospital. The method may yield sufficient accuracy for many tasks in the multi-centric setting and can easily be adapted to other languages/specialities.

A Study of the Spatial–Temporal Development Patterns and Influencing Factors of China’s National Archaeological Site Parks

National Archaeological Site Parks are strategic projects in China for innovatively exploring the scientific protection and rational utilization of large heritage sites for the purpose of balancing urban development and protecting cultural heritage. Therefore, research on the spatial–temporal evolution and influencing factors of National Archaeological Site Parks can improve and optimize their management and pattern systems and is therefore of great significance for the sustainable development of large sites and their surrounding areas. Targeting the 135 National Archaeological Site Parks of China, this paper adopts the nearest-neighbor index analytical method, the kernel density estimation method, the standard deviation ellipse method, the method of constructing an indicator system, and an SPSS factor analysis method to analyze the spatial–temporal evolution and influencing factors of these parks. The findings are as follows: (1) In terms of the temporal evolution, the number of National Archaeological Site Parks increased from 2010 to 2022, and the ages and types of the large heritage sites they are built upon gradually became more balanced and diverse. (2) In terms of the spatial evolution, National Archaeological Site Parks form two high-density areas in Central China and East China. Their overall distribution is consistent with China’s “Hu Line”. (3) In terms of spatial–temporal evolution, the center of gravity of National Archaeological Site Parks’ distribution is advancing toward Southwest China, and the trend of a more balanced distribution is rising. (4) Finally, regional development, heritage tourism, historical resources, and government support were observed to be factors that affect the spatial–temporal patterns of the National Archaeological Site Parks. Based on these findings, we propose specific strategies for coordinating and linking the above four major influencing factors to promote the rational utilization of large heritage sites and the sustainable development of National Archaeological Site Parks. We aim to improve and optimize the management and pattern systems of National Archaeological Site Parks, to promote urban renewal based on large heritage sites, and to provide valuable insights for policy makers and heritage practitioners in other countries with historical resources around the world.

Rank-based Hashing for Effective and Efficient Nearest Neighbor Search for Image Retrieval

The large and growing amount of digital data creates a pressing need for approaches capable of indexing and retrieving multimedia content. A traditional and fundamental challenge consists of effectively and efficiently performing nearest-neighbor searches. After decades of research, several different methods are available, including trees, hashing, and graph-based approaches. Most of the current methods exploit learning to hash approaches based on deep learning. In spite of effective results and compact codes obtained, such methods often require a significant amount of labeled data for training. Unsupervised approaches also rely on expensive training procedures usually based on a huge amount of data. In this work, we propose an unsupervised data-independent approach for nearest neighbor searches, which can be used with different features, including deep features trained by transfer learning. The method uses a rank-based formulation and exploits a hashing approach for efficient ranked list computation at query time. A comprehensive experimental evaluation was conducted on 7 public datasets, considering deep features based on CNNs and Transformers. Both effectiveness and efficiency aspects were evaluated. The proposed approach achieves remarkable results in comparison to traditional and state-of-the-art methods. Hence, it is an attractive and innovative solution, especially when costly training procedures need to be avoided.

METHODS OF KINSHIP ANALYSIS IN FOSSIL POPULATIONS

Introduction. Identifying kinship relationships from skeletal remains is among the various objectives of bioarchaeological studies. This article focuses on reviewing the methods used to analyze biological kinship in human fossil populations through non-metric traits. Methods. Since direct molecular-genetic analysis of kinship is often highly challenging due to the poor preservation of ancient DNA, special attention in such studies is given to nonmetric phenotypic traits. Results. Research with osteological samples that have been documented provides compelling evidence that the level of morphological similarity between individuals is directly related to their degree of biological kinship. In cases where the pedigrees of osteological materials are fully or partially known, phenotypic data can be effectively used in lieu of genetic information. Discussion. The methodology developed for kinship analysis depends on the internal spatial structure of the cemetery being studied. When analyzing small burial sites, the aim is to determine if the people buried there are close relatives. Various methods are used in these analyses, including different techniques for determining the likelihood of kinship, cluster analysis, and correlation coefficients. Identifying kinship is most promising in burial sites where archaeological or historical indicators of biological relationships are present. Kinship analysis in spatially structured cemeteries is aimed at identifying families or social groups. The analysis of uniformly distributed cemeteries focuses on identifying closely related individuals in large burials without clearly defined subgroups. This involves spatial correlation analysis, which tests for significant correlation between the matrix of spatial distances and the matrix of phenotypic distances; various counting methods to test for non-random clustering of traits; the nearest neighbor method; and a non-spatial block search procedure that simultaneously identifies presumed relatives and the traits that indicate the degree of their kinship. Conclusion. Many problems in establishing kinship can be overcome with the availability of skeletal material accompanied by verified genealogical data. Unfortunately, skeletal remains with preserved documentation are quite rare, limiting the opportunities to study the inheritance of non-metric traits and the morphological similarity of biologically related individuals.

FLEX: A fast and light-weight learned index for kNN search in high-dimensional space

The k Nearest Neighbors (kNN) search in high-dimensional space is a fundamental problem with various applications. In this paper, we try to solve this problem by using deep neural networks (DNNs). We apply DNNs to represent complex correlations between high-dimensional objects, enabling us to project similar objects into the same class, thus reducing the search cost. Based on DNNs, we propose two novel techniques to improve query efficiency while keeping accuracy. First, traditional DNNs typically demand extensive training data for achieving high accuracy. To decrease the training size, we design a multi-module DNN framework comprising several small modules. Each module learns to capture part of knowledge for the given query. The collective output of these sub-modules is then seamlessly integrated to form the final result. Second, with machine learning models, the size of candidates are unbounded. Thus, we design a linear-time data layout refinement algorithm, aiming to limit the number of candidates to a small constant. Empirically we find that our approach significantly outperforms the state-of-the-art methods in terms of both time efficiency and space efficiency while still attaining comparable or better accuracy.

Optimizing Routes for Sustainability: A Comparative Analysis of Parcel Distribution Methods in South Jakarta

This study focuses on improving the efficiency of PT. XYZ’s logistics services by examining the optimization of parcel distribution routes in South Jakarta. It compares the Nearest Neighbor Method and the Clarke and Wright Savings Method to determine the optimal strategy for delivering 53 packages to 8 locations. The Nearest Neighbor Method generates two routes, one encompassing 32 packages across four locations and the other encompassing 21 packages across four locations. This method significantly reduces travel distance, saving 27,3 kilometers in comparison to actual courier routes. The Clarke and Wright Savings Method generates two routes, one covering 35 packages across six locations and the other covering 18 packages across two locations. Compared to the actual van courier route, this method reduces the distance by 24,8 kilometers. The Nearest Neighbor Method reduces travel distances by 27.3 kilometers compared to actual routes, surpassing the Clarke and Wright Savings Method by 2.5 kilometers. Reduced travel distances reduce fuel consumption and greenhouse gas emissions, making logistics operations more sustainable on an environmental and economic level. This study concludes that the Nearest Neighbor Method is the optimal strategy for improving parcel distribution efficiency in South Jakarta, thereby substantially reducing the travel distances of PT. XYZ’s logistics operations.

A close-packed sphere model for characterising porous networks in atomistic simulations and its application in energy storage and conversion

Hierarchical (micro, meso & macro) porosity in materials plays a crucial role in influencing the movement of ions which governs the energy and power density during energy storage and conversion. The extant available methods to characterise porosity across scales (nano to meso to macro) lacks rigour and accuracy. Having accurate assessment of the porosity in materials can unlock new designs of electrodes for energy efficient energy storage and conversion devices such as batteries, supercapacitors and fuel cells. Through this work, we report the systematic development of a method to fully characterise the carbon porous networks using a molecular dynamics simulation testbed. Our work entails modelling and simulation of porous carbon structures using quenched molecular dynamics (QMD) simulations using Gaussian Approximation potential (GAP) and benchmarking the results with prior literature. This modelling technique can reliably be used for quantitative characterisation of the interconnectivity in porous structures to study ionic movements and charge transfer mechanisms. A new parameter, namely nearest neighbour search (NNS) coefficient was introduced to quantify homogeneity and networking in the porous structures. NNS coefficient increased from 1.62 to 1.92 with decrease of the annealing temperature from 8000 K to 4000 K in carbon. The procedure outlined was although tested on porous carbon networks, but adaptable to study any other material system at multi-length scales.

A nearest neighbor query method for searching objects with time and location informations based on spatiotemporal similarity

A nearest neighbor query method for searching objects with time and location informations based on spatiotemporal similarity

Combining Multiple Supervision for Robust Zero-Shot Dense Retrieval

Recently, dense retrieval (DR) models, which represent queries and documents with fixed-width vectors and retrieve relevant ones via nearest neighbor search, have drawn increasing attention from the IR community. However, previous studies have shown that the effectiveness of DR critically relies on sufficient training signals, which leads to severe performance degradation when applied in out-of-domain scenarios, where large-scale training data are usually unavailable. To solve this problem, existing studies adopt a data-augmentation-plus-joint-training paradigm to construct weak/pseudo supervisions on the target domain and combine them with the large-scale human annotated data on the source domain to train the DR models. However, they don't explicitly distinguish the data and the supervision signals in the training process and simply assume that the DR models are mighty enough to capture and memorize different domain knowledge and relevance matching patterns without guidance, which, as shown in this paper, is not true. Based on this observation, we propose a Robust Multi-Supervision Combining strategy (RMSC) that decouples the domain and supervision signals by explicitly telling the DR models how the domain data and supervision signals are combined in the training data with specially designed soft tokens. With the extra soft tokens to store the domain-specific and supervision-specific knowledge, RMSC allows the DR models to conduct retrieval based on human-like relevance matching patterns and target-specific language distribution on the target domain without human annotations. Extensive experiments on zero-shot DR benchmarks show that RMSC significantly improves the ranking performance on the target domain compared to strong DR baselines and domain adaptation methods, while being stable during training and can be combined with query generation or second-stage pre-training.