Small Set Research Articles

Cadastral-to-Agricultural: A Study on the Feasibility of Using Cadastral Parcels for Agricultural Land Parcel Delineation

Agricultural land parcels (ALPs) are essential for effective agricultural management, influencing activities ranging from crop yield estimation to policy development. However, traditional methods of ALP delineation are often labor-intensive and require frequent updates due to the dynamic nature of agricultural practices. Additionally, the significant variations across different regions and the seasonality of agriculture pose challenges to the automatic generation of accurate and timely ALP labels for extensive areas. This study introduces the cadastral-to-agricultural (Cad2Ag) framework, a novel approach that utilizes cadastral data as training labels to train deep learning models for the delineation of ALPs. Cadastral parcels, which are relatively widely available and stable elements in land management, serve as proxies for ALP delineation. Employing an adapted U-Net model, the framework automates the segmentation process using remote sensing images and geographic information system (GIS) data. This research evaluates the effectiveness of the proposed Cad2Ag framework in two U.S. regions—Indiana and California—characterized by diverse agricultural conditions. Through rigorous evaluation across multiple scenarios, the study explores diverse scenarios to enhance the accuracy and efficiency of ALP delineation. Notably, the framework demonstrates effective ALP delineation across different geographic contexts through transfer learning when supplemented with a small set of clean labels, achieving an F1-score of 0.80 and an Intersection over Union (IoU) of 0.67 using only 200 clean label samples. The Cad2Ag framework’s ability to leverage automatically generated, extensive, free training labels presents a promising solution for efficient ALP delineation, thereby facilitating effective management of agricultural land.

Temporal Sequencing of Documents

We outline an unsupervised method for temporal rank ordering of sets of historical documents, namely American State of the Union Addresses and DEEDS, a corpus of medieval English property transfer documents. Our method relies upon effectively capturing the gradual change in word usage via a bandwidth estimate for the non-parametric Generalized Linear Models (Fan, Heckman, and Wand, 1995). The number of possible rank orders needed to search through for cost functions related to the bandwidth can be quite large, even for a small set of documents. We tackle this problem of combinatorial optimization using the Simulated Annealing algorithm, which allows us to obtain the optimal document temporal orders. Our rank ordering method significantly improved the temporal sequencing of both corpora compared to a randomly sequenced baseline. This unsupervised approach should enable the temporal ordering of undated document sets.

Controlling the scatterplot shapes of 2D and 3D multidimensional projections

Multidimensional projections are effective techniques for depicting high-dimensional data. The point patterns created by such techniques, or a technique’s visual signature, depend — apart from the data themselves — on the technique design and its parameter settings. Controlling such visual signatures — something that only few projections allow — can bring additional freedom for generating insightful depictions of the data. We present a novel projection technique — ShaRP — that allows explicit control on such visual signatures in terms of shapes of similar-value point clusters (settable to rectangles, triangles, ellipses, and convex polygons) and the projection space (2D or 3D Euclidean or S2). We show that ShaRP scales computationally well with dimensionality and dataset size, provides its signature-control by a small set of parameters, allows trading off projection quality to signature enforcement, and can be used to generate decision maps to explore the behavior of trained machine-learning classifiers.

Geometrical Penrose tilings are characterized by their 1-atlas

Penrose rhombus tilings are tilings of the plane by two decorated rhombi such that the decorations match at the junction between two tiles (like in a jigsaw puzzle). In dynamical terms, they form a tiling space of finite type. If we remove the decorations, we get, by definition, a sofic tiling space that we here call geometrical Penrose tilings. Here, we show how to compute the patterns of a given size which appear in these tilings by three different methods: two based on the substitutive structure of the Penrose tilings and the last on their definition by the cut and projection method. We use this to prove that the geometrical Penrose tilings are characterized by a small set of patterns called vertex-atlas, i.e., they form a tiling space of finite type. Though considered as folklore, no complete proof of this result has been published, to our knowledge.

Randomised Badger Culling Trial lacks evidence for proactive badger culling effect on tuberculosis in cattle: comment on Mills et al. 2024, Parts I & II.

Re-evaluation of statistical analysis of the Randomised Badger Culling Trial (RBCT) by Torgerson et al. 2024 was rebutted by Mills et al. 2024 Parts I and II. The rebuttal defended the use of count rather than rate when considering bovine tuberculosis herd incidence. The defence makes biologically implausible use of Information Criterion for appraisal diagnostics; overfits data; and has erroneous Bayesian analyses. It favours 'goodness of fit' over 'predictive power', for a small data set, when the study was to inform application. Importantly, for 'total' bTB breakdown: ('confirmed' (OTF-W) +'unconfirmed' (OTF-S)), where modern interpretation of the main diagnostic bTB test better indicates the incidence rate of herd breakdown, there is no effect in cull and neighbouring areas, across all statistical models. The RBCT was a small, single experiment with unknown factors. With respect to the paradigm of reproducibility and the FAIR principles, the original RBCT analysis and recent efforts to support it are wholly unconvincing. The 2006 conclusion of the RBCT that "badger culling is unlikely to contribute positively to the control of cattle TB in Britain" is supported, but the route to such a position is revised in the light of modern veterinary understanding and statistical reappraisal.

Pottery from Motion – A Refined Approach to the Large-Scale Documentation of Pottery Using Structure from Motion

Abstract In recent years, interest in the usage of computer-based methods in archaeology, especially regarding field documentation, has grown significantly. In 2021, Göttlich et al. presented a new large-scale three-dimensional (3D) capture method for the documentation of pottery using structure from motion. This method, however, was only tested on a very small sample set and never truly conducted in a large-scale documentation surrounding. Consequently, we decided to test this workflow on a large-scale basis during three field campaigns (March 2022, November 2022, and March 2023) in Lebanon, documenting more than 4,000 sherds in total. In this article, I will present the results and observations of these campaigns, critically discuss the workflow involved (documentation and processing), and propose a refined workflow for this methodology. This article focuses solely on the relevant documentation and 3D processing.

Genetic Programming for Automatically Evolving Multiple Features to Classification

Abstract Performing classification on high-dimensional data poses a significant challenge due to the huge search space. Moreover, complex feature interactions introduce an additional obstacle. The problems can be addressed by using feature selection to select relevant features or feature construction to construct a small set of high-level features. However, performing feature selection or feature construction only might make the feature set suboptimal. To remedy this problem, this study investigates the use of genetic programming for simultaneous feature selection and feature construction in addressing different classification tasks. The proposed approach is tested on 16 datasets and compared with seven methods including both feature selection and feature constructions techniques. The results show that the obtained feature sets with the constructed and/or selected features can significantly increase the classification accuracy and reduce the dimensionality of the datasets. Further analysis reveals the complementarity of the obtained features leading to the promising classification performance of the proposed method.

Advanced Model Consistency Restoration with Higher-Order Short-Cut Rules

Sequential model synchronisation is the task of propagating changes from one model to another correlated one to restore consistency. It is challenging to perform this propagation in a least-changing way that avoids unnecessary deletions (which might cause information loss). From a theoretical point of view, so-called short-cut (SC) rules have been developed that enable provably correct propagation of changes while avoiding information loss. However, to be able to react to every possible change, an infinite set of such rules might be necessary. Practically, only small sets of pre-computed basic SC rules have been used, severely restricting the kind of changes that can be propagated without loss of information. In this work, we close that gap by developing an approach to compute more complex required SC rules on-the-fly during synchronisation. These higher-order SC rules allow us to cope with more complex scenarios when multiple changes must be handled in one step. We implemented our approach in the model transformation tool eMoflon. An evaluation shows that the overhead of computing higher-order SC rules on-the-fly is tolerable and at times even improves the overall performance. Above that, completely new scenarios can be dealt with without the loss of information.

Except-Extremes Ranked Set Sampling for Estimating the Population Variance with Two Applications of Real Data Sets

The ranked set sampling (RSS) procedure was initially established by McIntyre (1952) for estimating the mean of forage and pasture yield as more precise than simple random sampling (SRS). Recently, Aldrabseh and Ismail (2023) suggested the except extreme RSS (EERSS) approach as a modification to RSS for estimating the population mean. In this paper, a new estimator of the population variance is proposed using the EERSS method. The mean squared error and bias equations of the new estimator are derived. When the underlying distribution is non-symmetric, a simulation study is conducted to evaluate the suggested estimator relative to SRS and RSS, based on the same number of measured units, in terms of the relative precision and bias values for several sample sizes. For symmetric distributions, the exact values of the bias and relative precision of the EERSS variance estimator are evaluated. Two real datasets are utilized to illustrate the performance of the suggested variance estimator. It is found that the EERSS variance estimator is more efficient than the SRS estimator and more precise than RSS in most cases, especially for small set sizes.

Individual participant data meta-analysis to examine linear or non-linear treatment-covariate interactions at multiple time-points for a continuous outcome.

Individual participant data (IPD) meta-analysis projects obtain, harmonise, and synthesise original data from multiple studies. Many IPD meta-analyses of randomised trials are initiated to identify treatment effect modifiers at the individual level, thus requiring statistical modelling of interactions between treatment effect and participant-level covariates. Using a two-stage approach, the interaction is estimated in each trial separately and combined in a meta-analysis. In practice, two complications often arise with continuous outcomes: examining non-linear relationships for continuous covariates and dealing with multiple time-points. We propose a two-stage multivariate IPD meta-analysis approach that summarises non-linear treatment-covariate interaction functions at multiple time-points for continuous outcomes. A set-up phase is required to identify a small set of time-points; relevant knot positions for a spline function, at identical locations in each trial; and a common reference group for each covariate. Crucially, the multivariate approach can include participants or trials with missing outcomes at some time-points. In the first stage, restricted cubic spline functions are fitted and their interaction with each discrete time-point is estimated in each trial separately. In the second stage, the parameter estimates defining these multiple interaction functions are jointly synthesised in a multivariate random-effects meta-analysis model accounting for within-trial and across-trial correlation. These meta-analysis estimates define the summary non-linear interactions at each time-point, which can be displayed graphically alongside confidence intervals. The approach is illustrated using an IPD meta-analysis examining effect modifiers for exercise interventions in osteoarthritis, which shows evidence of non-linear relationships and small gains in precision by analysing all time-points jointly.

Effect of Data Quality and Data Quantity on the Estimation of Intrinsic Solubility: Analysis Based on a Single-Source Data Set.

Aqueous solubility is one of the most important physicochemical properties of drug molecules and a major driving force for oral drug absorption. To date, the performance of in silico models for the estimation of solubility for novel chemical space is limited. To investigate possible reasons and remedies for this, the Johnson and Johnson in-house aqueous solubility data with over 40,000 compounds was leveraged. All data were generated through the same high-throughput assay, providing a unique opportunity to explore the relationship between data quality, quantity, and model estimations. Six intrinsic solubility data sets with different sizes and noise levels were generated by making use of three different approaches: (i) inclusion or exclusion of amorphous solid residue, (ii) measured or experimental log D to identify the intrinsic solubility, and (iii) adopting or omitting a quality check process in the data processing workflow. A random forest regressor was trained on the data sets with three different sets of descriptors calculated from RDKit, ADMET predictor, or Mordred, and the performances were evaluated with nested cross-validation as well as ten refined test sets. The models confirm, as expected, that with the same data set size, high-quality data leads to better model performance; however, also, models trained with larger data sets containing analytical variability can give equally accurate estimations compared to models trained with small, clean, and diverse data sets. However, noise introduced by including the presence of amorphous solid postsolubility measurement in the training data set cannot be overcome by increasing data size, as they are introducing a biased systematic positive error in the data set, confirming the importance of critical data review. Finally, two top-performing models were tested on the first test set from the second solubility challenge, achieving RMSE values of 0.74 and 0.72 and log S ± 0.5 of 46 and 48%, respectively. These results demonstrated improved performance compared to those reported in the findings of the competition, highlighting that a single-source curated data set can enhance the prediction of intrinsic solubility.

NeRF-OR: neural radiance fields for operating room scene reconstruction from sparse-view RGB-D videos.

RGB-D cameras in the operating room (OR) provide synchronized views of complex surgical scenes. Assimilation of this multi-view data into a unified representation allows for downstream tasks such as object detection and tracking, pose estimation, and action recognition. Neural radiance fields (NeRFs) can provide continuous representations of complex scenes with limited memory footprint. However, existing NeRF methods perform poorly in real-world OR settings, where a small set of cameras capture the room from entirely different vantage points. In this work, we propose NeRF-OR, a method for 3D reconstruction of dynamic surgical scenes in the OR. Where other methods for sparse-view datasets use either time-of-flight sensor depth or dense depth estimated from color images, NeRF-OR uses a combination of both. The depth estimations mitigate the missing values that occur in sensor depth images due to reflective materials and object boundaries. We propose to supervise with surface normals calculated from the estimated depths, because these are largely scale invariant. We fit NeRF-OR to static surgical scenes in the 4D-OR dataset and show that its representations are geometrically accurate, where state of the art collapses to sub-optimal solutions. Compared to earlier work, NeRF-OR grasps fine scene details while training 30 faster. Additionally, NeRF-OR can capture whole-surgery videos while synthesizing views at intermediate time values with an average PSNR of 24.86dB. Last, we find that our approach has merit in sparse-view settings beyond those in the OR, by benchmarking on the NVS-RGBD dataset that contains as few as three training views. NeRF-OR synthesizes images with a PSNR of 26.72dB, a 1.7% improvement over state of the art. Our results show that NeRF-OR allows for novel view synthesis with videos captured by a small number of cameras with entirely different vantage points, which is the typical camera setting in the OR. Code is available via: github.com/Beerend/NeRF-OR .

Bundled matrix product states represent low-energy excitations faithfully

Abstract We consider a set of density matrices. All of which are written in the same orbital basis, but the orbital basis size is less than the total Hilbert space size. We ask how each density matrix is related to each of the others by establishing a norm between density matrices based on the truncation error in a partial trace for a small set of orbitals. We find that states with large energy differences must have large differences in their density matrices. Small energy differences are divided into two groups, one where two density matrices have small differences and another where they are very different, as is the case of symmetry. We extend these ideas to a bundle of matrix product states and show that bond dimension of the wavefunction ansatz for two states with large energy differences are larger. Meanwhile, low energy differences can have nearly the same bond dimensions for similar states.

A Machine Learning Method for RNA-Small Molecule Binding Preference Prediction.

The interaction between RNA and small molecules is crucial in various biological functions. Identifying molecules targeting RNA is essential for the inhibitor design and RNA-related studies. However, traditional methods focus on learning RNA sequence and secondary structure features and neglect small molecule characteristics, and resulting in poor performance on unknown small molecule testing. To overcome this limitation, we developed a double-layer stacking-based machine learning model called ZHMol-RLinter. This approach more effectively predicts RNA-small molecule binding preferences by learning RNA and small molecule features to capture their interaction information. ZHMol-RLinter also combines sequence and secondary structural features with structural geometric and physicochemical environment information to capture the specificity of RNA spatial conformations in recognizing small molecules. Our results demonstrate that ZHMol-RLinter has a success rate of 90.8% on the published RL98 testing set, representing a significant improvement over existing methods. Additionally, ZHMol-RLinter achieved a success rate of 77.1% on the unknown small molecule UNK96 testing set, showing substantial improvement over the existing methods. The evaluation of predicted structures confirms that ZHMol-RLinter is reliable and accurate for predicting RNA-small molecule binding preferences, even for challenging unknown small molecule testing. Predicting RNA-small molecule binding preferences can help in the understanding of RNA-small molecule interactions and promote the design of RNA-related drugs for biological and medical applications.

Automated design of multi-target ligands by generative deep learning

Generative deep learning models enable data-driven de novo design of molecules with tailored features. Chemical language models (CLM) trained on string representations of molecules such as SMILES have been successfully employed to design new chemical entities with experimentally confirmed activity on intended targets. Here, we probe the application of CLM to generate multi-target ligands for designed polypharmacology. We capitalize on the ability of CLM to learn from small fine-tuning sets of molecules and successfully bias the model towards designing drug-like molecules with similarity to known ligands of target pairs of interest. Designs obtained from CLM after pooled fine-tuning are predicted active on both proteins of interest and comprise pharmacophore elements of ligands for both targets in one molecule. Synthesis and testing of twelve computationally favored CLM designs for six target pairs reveals modulation of at least one intended protein by all selected designs with up to double-digit nanomolar potency and confirms seven compounds as designed dual ligands. These results corroborate CLM for multi-target de novo design as source of innovation in drug discovery.

Replicable and robust cellular and biochemical blood marker signatures of depression and depressive symptoms

Background: Identification of replicable and robust peripheral blood-based markers associated with depression remains elusive, given that studies frequently identify potential biomarkers that ultimately fail to replicate in other studies, impeding progress in psychiatric research. Peripheral biochemical and cellular markers (PBCs; e.g., albumin) may play an important role in depression. Methods: Using a test-replication design including participants from the NHANES community cohort (ntest=17,450, nreplication=17,449), we examined 42 PBCs to identify PBCs that were both replicably and robustly associated with either overall depression severity or individual symptoms of depression across both cohorts across a wide range of possible combinations of analytic decisions (n's = 17,000+). Results: We found that a small set of PBCs (e.g., bilirubin) were robustly and replicably associated with overall depression severity, with unique signatures of PBCs linked with individual symptoms of depression when stratified by gender. A varying degree of correlation was found between measures of replicability. Conclusions: We identified replicable and robust cellular biochemical blood marker signatures associated with both overall depression severity and individual symptoms of depression. Our findings can be used to enhance other researchers’ abilities to better understand factors associated with depression and potentially drive the development of effective treatments for depression.

Interactions of Potential Endocrine-Disrupting Chemicals with Whole Human Proteome Predicted by AlphaFold2 Using an In Silico Approach.

Binding with proteins is a critical molecular initiating event through which environmental pollutants exert toxic effects in humans. Previous studies have been limited by the availability of three-dimensional (3D) protein structures and have focused on only a small set of environmental contaminants. Using the highly accurate 3D protein structure predicted by AlphaFold2, this study explored over 60 million interactions obtained through molecular docking between 20,503 human proteins and 1251 potential endocrine-disrupting chemicals. A total of 66,613,773 docking results were obtained, 1.2% of which were considered to be high binding, as their docking scores were lower than -7. Monocyte to macrophage differentiation factor 2 (MMD2) was predicted to interact with the highest number of environmental pollutants (526), with polychlorinated biphenyls and polychlorinated dibenzofurans accounting for a significant proportion. Dimension reduction and clustering analysis revealed distinct protein profiles characterized by high binding affinities for perfluoroalkyl and polyfluoroalkyl substances (PFAS), phthalate-like chemicals, and other pollutants, consistent with their uniquely enriched pathways. Further structural analysis indicated that binding pockets with a high proportion of charged amino acid residues, relatively low α-helix content, and high β-sheet content were more likely to bind to PFAS than others. This study provides insights into the toxicity pathways of various pollutants impacting human health and offers novel perspectives for the establishment and expansion of adverse outcome pathway-based models.

Comparing methods for estimating causal treatment effects of administrative health data: A plasmode simulation study.

Estimating the causal effects of health policy interventions is crucial for policymaking but is challenging when using real-world administrative health care data due to a lack of methodological guidance. To help fill this gap, we conducted a plasmode simulation using such data from a recent policy initiative launched in a deprived urban area in Germany. Our aim was to evaluate and compare the following methods for estimating causal effects: propensity score matching, inverse probability of treatment weighting, and entropy balancing, all combined with difference-in-differences analysis, augmented inverse probability weighting, and targeted maximum likelihood estimation. Additionally, we estimated nuisance parameters using regression models and an ensemble learner called superlearner. We focused on treatment effects related to the number of physician visits, total health care cost, and hospitalization. While each approach has its strengths and weaknesses, our results demonstrate that the superlearner generally worked well for handling nuisance terms in large covariate sets when combined with doubly robust estimation methods to estimate the causal contrast of interest. In contrast, regression-based nuisance parameter estimation worked best in small covariate sets when combined with singly robust methods.

Assessing the implementation of a tertiary care comprehensive pediatric asthma education program using electronic medical records and decision support tools

Background Self-management education is integral for proper asthma management. However, there is an accessibility gap to self-management education following asthma hospitalizations. Most pediatric patients and their families receive suboptimal or no education. Objective To implement a comprehensive pediatric asthma education program and evaluate subsequent self-management knowledge in patients as well as behavior change outcomes reflected in the frequency of asthma related repeat emergency department visits and hospitalization. The program implementation was informed by the Knowledge to Translation Action Framework and the i-PARIHS model for quality improvement and involved several iterative stages. Methods We implemented a comprehensive asthma education program for the families of all children 0-18 years old who had been admitted for an asthma exacerbation to the Children’s Hospital of Eastern Ontario (CHEO), beginning on April 1, 2018. The program was adapted to the stages of the Knowledge Translation to Action Framework including undertaking an environmental scan, expert stakeholder feedback, reviews, addressing barriers, and tailoring the intervention, along with evaluating knowledge and health outcomes. Education was delivered over 1-2 h in personalized individual or small group settings, within 4 wk of hospital discharge. All education was provided by registered nurses or respiratory therapists who were also certified asthma educators. The EPIC electronic medical record was used to facilitate referral and scheduling of asthma education sessions, and to track subsequent acute asthma visits. We compared the frequency of a repeat asthma emergency department (ED) visit or hospitalization within 1-year following an initial asthma hospitalization for children who would have received comprehensive asthma education, to a historical cohort of children who were hospitalized between April 9, 2017 – Apr 8, 2018, and did not receive asthma education. Results The program had a high enrollment, capturing nearly 75% of the target population. Most families found the program to be acceptable and reported increased knowledge of how to manage asthma. We identified a crude overall 54% reduction in repeat hospitalizations among children 1 year after implementation of the asthma education program (i.e. 10.2% (23/225) repeat hospitalization rate pre- implementation versus 4.8% (11/227) post-implementation). In adjusted time-to event analysis, this reduction was prominent at 3 months among those who received comprehensive asthma education, relative to those who did not, but this improvement was not sustained by 1 year (HR =1.1, 95% CI =0.55- 2.05; p-value = 0.6). Discussion Although we did not find long-term improvements in ED visits, or hospitalizations, in children of caregivers who participated in comprehensive asthma education, the asthma education program holds potential given that most patients found it to be acceptable and that it increased asthma management knowledge. A future asthma education program should include multiple sessions to ensure that the knowledge and behavior change will be sustained, leading ultimately to long-term reductions in repeat ED visits and hospitalizations.

Cross-lingual dependency parsing for a language with a unique script

Abstract Syntactic parsing is one of the areas in Natural Language Processing. The development of large-scale multilingual language models has enabled cross-lingual parsing approaches, which allows us to develop parsers for languages that do not have treebanks available. However, these approaches rely on the assumption that languages share orthographic representations and lexical entries. In this article, we investigate methods for developing a dependency parser for Xibe, a low-resource language that is written in a unique script. We first investigate lexicalized monolingual dependency parsing experiments to examine the effectiveness of word, part-of-speech, and character embeddings as well as pre-trained language models. Results show that character embeddings can significantly improve performance, while pre-trained language models decrease performance since they do not recognize the Xibe script. We also train delexicalized monolingual models, which yield competitive results to the best lexicalized model. Since the monolingual models are trained on a very small training set, we also investigate lexicalized and delexicalized cross-lingual models. We use six closely related languages as source language, which cover a wide range of scripts. In this setting, the delexicalized models achieve higher performance than lexicalized models. A final experiment shows that we can increase performance of the cross-lingual model by combining source languages and selecting the most similar sentences to Xibe as training set. However, all cross-lingual parsing results are still considerably lower than the monolingual model. We attribute the low performance of cross-lingual methods to syntactic and annotation differences as well as to the impoverished input of Universal Dependency Part-of-Speech tags that the delexicalized model has access to.