Accurate Inference Research Articles

Lipids, lipid-lowering drugs and lateral epicondylitis of the humerus: a drug-targeted Mendelian randomization study.

Clinical observations indicate that blood lipids may be risk factors for lateral epicondylitis (LE) of the humerus, and lipid-lowering drugs are also used for the prevention and treatment of tendon diseases, but these lack high-quality clinical trial evidence and remain inconclusive. Mendelian randomization (MR) analyses can overcome biases in traditional observational studies and offer more accurate inference of causal relationships. Therefore, we employed this approach to investigate whether blood lipids are risk factors for LE and if lipid-lowering drugs can prevent it. Genetic variations associated with lipid traits, including low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), and total cholesterol (TC), were obtained from the UK Biobank and the Global Lipids Genetics Consortium (GLGC). Data on genetic variation in LE were sourced from FinnGen, including 24,061 patients and 275,212 controls. Subsequently, MR analyses were conducted to assess the potential correlation between lipid traits and LE. Additionally, drug-target Mendelian randomization analyses were performed on 10 drug targets relevant to LE. For those drug targets that yielded significant results, further analysis was conducted using colocalization techniques. No correlation was found between three blood lipid traits and LE. Lipoprotein lipase (LPL) enhancement is significantly associated with a decreased risk of LE (OR = 0.76, [95% CI, 0.65-0.90], p = 0.001). The expression of LPL in the blood is associated with LE and shares a single causal variant (12.07%), greatly exceeding the probability of different causal variations (1.93%), with a colocalization probability of 86.2%. The three lipid traits are not risk factors for lateral epicondylitis. LPL is a potential drug target for the prevention and treatment of LE.

GLAMLE: inference for multiview network data in the presence of latent variables, with an application to commodities trading

The statistical analysis of multiview network data provides researchers with insights on the mechanism governing many complex systems of interconnected elements. A graph is routinely applied to model this type of data: this poses several statistical challenges. Among them, there is the need for considering that some unobservable factors may determine the network connections and the graph topology. This requirement entails both complex models and computationally heavy inference procedures. To cope with these inference issues, a novel and flexible class of latent variable models labeled Graph Generalized Linear Latent Variable Model (GGLLVM) is introduced. Then, its parameters are estimated via the maximization of the Laplace-approximated likelihood. This leads to the creation of an M-estimator which is named Graph Laplace-Approximated Maximum Likelihood Estimator (GLAMLE). The statistical properties of this new estimator are characterized, and numerical experiments on simulated networks are performed to illustrate that the GLAMLE yields fast and accurate inference, in comparison to other widely used approaches. A real data application to commodities trading in European countries illustrates how the use of GGLLVM and GLAMLE unveils the import/export propensity that each node of the network has toward other nodes, along with additional information specific to each traded commodity.

256‐level honey memristor‐based in‐memory neuromorphic system

AbstractPromising synaptic behaviour has been exhibited by memristors based on natural organic materials. Such memristor‐based neuromorphic systems offer notable benefits, including environmental sustainability, low production and disposal costs, non‐volatile storage capability, and bio/Complementary Metal‐Oxide‐Semiconductor (CMOS) compatibility. Here, a 256‐level honey memristor‐based neuromorphic system is experimentally evaluated for image recognition. In detail, first, 256‐level honey memristors are manufactured and tested based on in‐house technology; next, the non‐linear characteristics and inherent variation of honey memristor devices, which lead to imprecise weight updates and limit the inference accuracy, are investigated. Experimental results indicate that the inference accuracy of the 256‐level honey memristor‐based neuromorphic system is greater than 88% without cycle‐to‐cycle variations and 87% with cycle‐to‐cycle variations for different optimization algorithms. The overall performance of optimization algorithms with and without variation is compared in terms of energy and latency, where the momentum algorithm consistently outperforms the rest of the algorithms. This 256‐level honey memristor is a promising alternative enabling sustainable neuromorphic systems, encouraging further research into natural organic materials for neuromorphic computing.

Scalable CNN-based classification of selective sweeps using derived allele frequencies.

Selective sweeps can successfully be distinguished from neutral genetic data using summary statistics and likelihood-based methods that analyze single nucleotide polymorphisms (SNPs). However, these methods are sensitive to confounding factors, such as severe population bottlenecks and old migration. By virtue of machine learning, and specifically convolutional neural networks (CNNs), new accurate classification models that are robust to confounding factors have been recently proposed. However, such methods are more computationally expensive than summary-statistic-based ones, yielding them impractical for processing large-scale genomic data. Moreover, SNP data are frequently preprocessed to improve classification accuracy, further exacerbating the long analysis times. To this end, we propose a 1D CNN-based model, dubbed FAST-NN, that does not require any preprocessing while using only derived allele frequencies instead of summary statistics or raw SNP data, thereby yielding a sample-size-invariant, scalable solution. We evaluated several data fusion approaches to account for the variance of the density of genetic diversity across genomic regions (a selective sweep signature), and performed an extensive neural architecture search based on a state-of-the-art reference network architecture (SweepNet). The resulting model, FAST-NN, outperforms the reference architecture by up to 12% inference accuracy over all challenging evolutionary scenarios with confounding factors that were evaluated. Moreover, FAST-NN is between 30× and 259× faster on a single CPU core, and between 2.0× and 6.2× faster on a GPU, when processing sample sizes between 128 and 1000 samples. Our work paves the way for the practical use of CNNs in large-scale selective sweep detection. https://github.com/SjoerdvandenBelt/FAST-NN.

Improving model robustness to weight noise via consistency regularization

Abstract As an emerging computing architecture, the computing-in-memory (CIM) exhibits significant potential for energy efficiency and computing power in artificial intelligence applications. However, the intrinsic non-idealities of CIM devices, manifesting as random interference on the weights of neural network, may significantly impact the inference accuracy. In this paper, we propose a novel training algorithm designed to mitigate the impact of weight noise. The algorithm strategically minimizes cross-entropy loss while concurrently refining the feature representations in intermediate layers to emulate those of an ideal, noise-free network. This dual-objective approach not only preserves the accuracy of the neural network but also enhances its robustness against noise-induced degradation. Empirical validation across several benchmark datasets confirms that our algorithm sets a new benchmark for accuracy in CIM-enabled neural network applications. Compared to the most commonly used forward noise training methods, our approach yields approximately a 2% accuracy boost on the ResNet32 model with the CIFAR-10 dataset and a weight noise scale of 0.2, and achieves a minimum performance gain of 1% on ResNet18 with the ImageNet dataset under the same noise quantization conditions.

Defying the flat-Earth myth: Soil organic carbon monitoring in topographically complex temperate pasture landscapes

Soil sampling design to capture the spatiotemporal variability of soil organic carbon (SOC) stocks in complex landscapes is challenged by the influence of topographical features on primary production, animal behaviour, and carbon and nutrient return to grasslands. Here we explore a range of approaches for streamlining the sampling process using measures of SOC stock data (2003, 2020, 2021) from a long-term hill-country fertilizer experiment. The stepwise procedure included a) evaluation of sampling efficiency between stratified and random soil sampling designs to capture spatial variability of SOC stocks, b) semi-parametric regression analysis to evaluate the representativeness of certain strata (e.g. medium slopes), c) exploring the appropriateness of using shallow SOC stock estimates to represent the full sampling depth, and d) the use of a process-based model to determine expected SOC stock changes and temporal variability to estimate the number of representative soil samples required to achieve a significant minimum change. Our analyses suggest that a) the stratified sampling design was more efficient reflected in the number of samples required to identify 5 % deviation in SOC stocks, b) estimates obtained on the medium slope were a reasonable approximation of the SOC stocks and its spatial variation across the farmlets, c) shallow (0–75 mm) SOC stock estimates from an adjusted regression model offered an accurate inference of its vertical distribution (0–300 mm), and d) the Grass-NEXT model provided insights on where more intensive sampling is required to detect a significant minimum change in SOC stocks over time. Tailored, stratified soil sampling and context-specific strategies, such as using representative strata and shallow sampling as effective steps for estimating SOC stocks and addressing data gaps, are crucial for simplifying monitoring and minimizing the cost of on-farm soil sampling designs in topographically complex landscapes covered by long-term pastures.

Poorer representation of minds underpins less accurate mental state inference for out-groups

Societies are becoming more polarised, driven in part by misconceptions about out-groups’ beliefs. To understand these effects, one must examine the cognitive processes underlying how people think about others. Here, we investigate whether people are less prone to theorise about the minds of out-groups, or less able to do so. Participants (Study 1: n = 128; Study 2: n = 128) made inferences about social and political beliefs held by real in-group and out-group members, and could choose to receive further information to improve these inferences. Results show: (1) participants sought equivalent or greater information about out-groups relative to in-groups; but despite this, (2) made significantly less accurate inferences for out-groups; and (3) were significantly less aware of their reduced ability. This shows that poorer mental state inference is not underpinned by a reduced propensity to consider out-group minds, but instead by a worse representation of the minds of out-groups.

Byzantine-robust Federated Learning via Cosine Similarity Aggregation

Federated Learning (FL) is proposed to train a machine learning model for clients with different training data. During the training of FL, a centralized server is usually employed to aggregate local models from clients iteratively. The aggregation process suffers from Byzantine attacks, where clients’ models could be maliciously modified by attackers to degrade the training performance. Existing defense aggregation solutions use distances or angles between different gradients to identify and eliminate malicious models from clients. However, they do not work well due to the high dimensional property of the machine learning model. Distance-based solutions cannot effectively identify attackers when the gradient direction of the model is maliciously tampered with. Angle-based solutions face the issue of low model accuracy for large models. In this paper, we propose Convolutional Kernel Angle-based Defense Aggregation (CKADA) to improve defense performance under various Byzantine attacks. The key of CKADA is to use the angle between convolutional kernels as the attack detection metric because the obtuse angle indicates the wrong training direction. CKADA calculates the angle between a client’s convolutional kernel gradients and the server’s convolutional kernel gradients as the attacker detection metric and eliminates convolutional kernel gradients of clients that create an obtuse angle to mitigate the impact of attackers on the model. We evaluate the performance of CKADA using AlexNet, ResNet-50, and GoogLeNet under two typical attacks. Simulation results show that CKADA mitigates the impact of Byzantine attacks and outperforms existing angle-based solutions and distance-based solutions by improving inference accuracy up to 67% and 89% respectively.

Beehive Smart Detector Device for the Detection of Critical Conditions That Utilize Edge Device Computations and Deep Learning Inferences.

This paper presents a new edge detection process implemented in an embedded IoT device called Bee Smart Detection node to detect catastrophic apiary events. Such events include swarming, queen loss, and the detection of Colony Collapse Disorder (CCD) conditions. Two deep learning sub-processes are used for this purpose. The first uses a fuzzy multi-layered neural network of variable depths called fuzzy-stranded-NN to detect CCD conditions based on temperature and humidity measurements inside the beehive. The second utilizes a deep learning CNN model to detect swarming and queen loss cases based on sound recordings. The proposed processes have been implemented into autonomous Bee Smart Detection IoT devices that transmit their measurements and the detection results to the cloud over Wi-Fi. The BeeSD devices have been tested for easy-to-use functionality, autonomous operation, deep learning model inference accuracy, and inference execution speeds. The author presents the experimental results of the fuzzy-stranded-NN model for detecting critical conditions and deep learning CNN models for detecting swarming and queen loss. From the presented experimental results, the stranded-NN achieved accuracy results up to 95%, while the ResNet-50 model presented accuracy results up to 99% for detecting swarming or queen loss events. The ResNet-18 model is also the fastest inference speed replacement of the ResNet-50 model, achieving up to 93% accuracy results. Finally, cross-comparison of the deep learning models with machine learning ones shows that deep learning models can provide at least 3-5% better accuracy results.

Infants evaluate informativeness of evidence and predict causal events as revealed in theta oscillations and predictive looking

This study investigates 16-month-old infants’ sensitivity to the informativeness of evidence and its potential link to infants’ ability to draw accurate causal inferences and predict unfolding events. Employing concurrent EEG and eye tracking, data from 66 infants revealed significantly increased theta oscillatory activity when infants expected to see causally unconfounded evidence compared to confounded evidence, suggesting heightened cognitive engagement in anticipation of informative evidence. Crucially, this difference was more pronounced in the subset of infants who later made correct predictions, suggesting that they had correctly inferred the causal structure based on the evidence presented. This research sheds light on infants’ motivation to seek explanatory causal information, suggesting that even at 16 months, infants can strategically direct attention to situations conducive to acquiring informative evidence, potentially laying the groundwork for the impressive abilities of humans to rapidly acquire knowledge and develop causal theories of the world.

Anatomizing Deep Learning Inference in Web Browsers

Web applications have increasingly adopted Deep Learning (DL) through in-browser inference , wherein DL inference performs directly within Web browsers. The actual performance of in-browser inference and its impacts on the quality of experience ( QoE ) remain unexplored, and urgently require new QoE measurements beyond traditional ones, e.g., mainly focusing on page load time. To bridge this gap, we make the first comprehensive performance measurement of in-browser inference to date. Our approach proposes new metrics to measure in-browser inference: responsiveness, smoothness, and inference accuracy. Our extensive analysis involves 9 representative DL models across Web browsers of 50 popular PC devices and 20 mobile devices. The results reveal that in-browser inference exhibits a substantial latency gap, averaging 16.9 times slower on CPU and 4.9 times slower on GPU compared to native inference on PC devices. The gap on mobile CPU and mobile GPU is 15.8 times and 7.8 times, respectively. Furthermore, we identify contributing factors to such latency gap, including underutilized hardware instruction sets, inherent overhead in the runtime environment, resource contention within the browser, and inefficiencies in software libraries and GPU abstractions. Additionally, in-browser inference imposes significant memory demands, at times exceeding 334.6 times the size of the DL models themselves, partly attributable to suboptimal memory management. We also observe that in-browser inference leads to a significant 67.2% increase in the time it takes for GUI components to render within Web browsers, significantly affecting the overall user QoE of Web applications reliant on this technology.

Corun: Concurrent Inference and Continuous Training at the Edge for Cost-Efficient AI-Based Mobile Image Sensing.

Intelligent mobile image sensing powered by deep learning analyzes images captured by cameras from mobile devices, such as smartphones or smartwatches. It supports numerous mobile applications, such as image classification, face recognition, and camera scene detection. Unfortunately, mobile devices often lack the resources necessary for deep learning, leading to increased inference latency and rapid battery consumption. Moreover, the inference accuracy may decline over time due to potential data drift. To address these issues, we introduce a new cost-efficient framework, called Corun, designed to simultaneously handle multiple inference queries and continual model retraining/fine-tuning of a pre-trained model on a single commodity GPU in an edge server to significantly improve the inference throughput, upholding the inference accuracy. The scheduling method of Corun undertakes offline profiling to find the maximum number of concurrent inferences that can be executed along with a retraining job on a single GPU without incurring an out-of-memory error or significantly increasing the latency. Our evaluation verifies the cost-effectiveness of Corun. The inference throughput provided by Corun scales with the number of concurrent inference queries. However, the latency of inference queries and the length of a retraining epoch increase at substantially lower rates. By concurrently processing multiple inference and retraining tasks on one GPU instead of using a separate GPU for each task, Corun could reduce the number of GPUs and cost required to deploy mobile image sensing applications based on deep learning at the edge.

A robust operators’ cognitive workload recognition method based on denoising masked autoencoder

Identifying the cognitive workload of operators is crucial in complex human-automation collaboration systems. An excessive workload can lead to fatigue or accidents, while an insufficient workload may diminish situational awareness and efficiency. However, existing supervised learning-based methods for workload recognition are ineffective when dealing with imperfect input data, such as missing or noisy data, which is not practical in real applications. This study introduces a robust Electroencephalogram (EEG)-enabled cognitive workload recognition model using self-supervised learning. The proposed method, DMAEEG, combines the training strategies of denoising autoencoders and masked autoencoders, demonstrating strong robustness against noisy and incomplete data. More specifically, we adopt the temporal convolutional network and multi-head self-attention mechanisms as the backbone, effectively capturing both the temporal and spatial features from EEG. Extensive experiments are conducted to verify the effectiveness and robustness of the proposed method on an open dataset and a self-collected dataset. The results indicate that DMAEEG performs superior to other state-of-the-art across various evaluation metrics. Moreover, DMAEEG maintains high accuracy in workload inference even when EEG signals are corrupted with a high masking ratio or strong noises. This signifies its superiority in capturing robust intrinsic patterns from imperfect EEG data. The proposed method significantly contributes to decoding EEG signals for workload recognition in real-world applications, thereby enhancing the safety and reliability of human-automation interactions.

TEFLON: Thermally Efficient Dataflow-aware 3D NoC for Accelerating CNN Inferencing on Manycore PIM Architectures

Resistive random-access memory (ReRAM)-based processing-in-memory (PIM) architectures are used extensively to accelerate inferencing/training with convolutional neural networks (CNNs). Three-dimensional (3D) integration is an enabling technology to integrate many PIM cores on a single chip. In this work, we propose the design of a t hermally e fficient data flo w-aware monolithic 3D (M3D) N oC architecture referred to as TEFLON to accelerate CNN inferencing without creating any thermal bottlenecks. TEFLON reduces the Energy-Delay-Product (EDP) by 42%, 46%, and 45% on an average compared to a conventional 3D mesh NoC for systems with 36-, 64-, and 100-PIM cores, respectively. TEFLON reduces the peak chip temperature by 25 K and improves the inference accuracy by up to 11% compared to sole performance-optimized SFC-based counterpart for inferencing with diverse deep CNN models using CIFAR-10/100 datasets on a 3D system with 100-PIM cores.

Insertable Glucose Sensor Using a Compact and Cost-Effective Phosphorescence Lifetime Imager and Machine Learning.

Optical continuous glucose monitoring (CGM) systems are emerging for personalized glucose management owing to their lower cost and prolonged durability compared to conventional electrochemical CGMs. Here, we report a computational CGM system, which integrates a biocompatible phosphorescence-based insertable biosensor and a custom-designed phosphorescence lifetime imager (PLI). This compact and cost-effective PLI is designed to capture phosphorescence lifetime images of an insertable sensor through the skin, where the lifetime of the emitted phosphorescence signal is modulated by the local concentration of glucose. Because this phosphorescence signal has a very long lifetime compared to tissue autofluorescence or excitation leakage processes, it completely bypasses these noise sources by measuring the sensor emission over several tens of microseconds after the excitation light is turned off. The lifetime images acquired through the skin are processed by neural network-based models for misalignment-tolerant inference of glucose levels, accurately revealing normal, low (hypoglycemia) and high (hyperglycemia) concentration ranges. Using a 1 mm thick skin phantom mimicking the optical properties of human skin, we performed in vitro testing of the PLI using glucose-spiked samples, yielding 88.8% inference accuracy, also showing resilience to random and unknown misalignments within a lateral distance of ∼4.7 mm with respect to the position of the insertable sensor underneath the skin phantom. Furthermore, the PLI accurately identified larger lateral misalignments beyond 5 mm, prompting user intervention for realignment. The misalignment-resilient glucose concentration inference capability of this compact and cost-effective PLI makes it an appealing wearable diagnostics tool for real-time tracking of glucose and other biomarkers.

Skew-Normal Inflated Models: Mathematical Characterization and Applications to Medical Data with Excess of Zeros and Ones

The modeling of data involving proportions, confined to a unit interval, is crucial in diverse research fields. Such data, expressing part-to-whole relationships, span from the proportion of individuals affected by diseases to the allocation of resources in economic sectors and the survival rates of species in ecology. However, modeling these data and interpreting information obtained from them present challenges, particularly when there is high zero–one inflation at the extremes of the unit interval, which indicates the complete absence or full occurrence of a characteristic or event. This inflation limits traditional statistical models, which often fail to capture the underlying distribution, leading to biased or imprecise statistical inferences. To address these challenges, we propose and derive the skew-normal zero–one inflated (SNZOI) models, a novel class of asymmetric regression models specifically designed to accommodate zero–one inflation presented in the data. By integrating a continuous-discrete mixture distribution with covariates in both continuous and discrete parts, SNZOI models exhibit superior capability compared to traditional models when describing these complex data structures. The applicability and effectiveness of the proposed models are demonstrated through case studies, including the analysis of medical data. Precise modeling of inflated proportion data unveils insights representing advancements in the statistical analysis of such studies. The present investigation highlights the limitations of existing models and shows the potential of SNZOI models to provide more accurate and precise inferences in the presence of zero–one inflation.

Combining mutation and recombination statistics to infer clonal families in antibody repertoires.

B-cell repertoires are characterized by a diverse set of receptors of distinct specificities generated through two processes of somatic diversification: V(D)J recombination and somatic hypermutations. B-cell clonal families stem from the same V(D)J recombination event, but differ in their hypermutations. Clonal families identification is key to understanding B-cell repertoire function, evolution, and dynamics. We present HILARy (high-precision inference of lineages in antibody repertoires), an efficient, fast, and precise method to identify clonal families from single- or paired-chain repertoire sequencing datasets. HILARy combines probabilistic models that capture the receptor generation and selection statistics with adapted clustering methods to achieve consistently high inference accuracy. It automatically leverages the phylogenetic signal of shared mutations in difficult repertoire subsets. Exploiting the high sensitivity of the method, we find the statistics of evolutionary properties such as the site frequency spectrum and dN/dS ratio do not depend on the junction length. We also identify a broad range of selection pressures spanning two orders of magnitude.

PODI: A Private Object Detection Inference framework for autonomous vehicles

Object detection is essential for autonomous vehicles to perceive and understand their environment. The restricted storage and processing capacities of vehicles necessitate the outsourcing of object detection services. However, this may raise concerns regarding the privacy of the uploaded images. Although there have been some studies on privacy-preserving object detection networks, they either lack location privacy protection or involve excessive computational and communication overheads. To address this issue, we propose a private object detection inference framework (PODI), which is based on a Faster R-CNN and aims to protect both classification and location privacy. PODI employs additive secret sharing protocols to support collaborative computation between two edge servers. By using efficient protocols such as secure Maxpool, secure array access, and secure exponent, PODI significantly reduces computational and communication overheads. A theoretical analysis has confirmed the security, correctness, and efficiency of PODI. Extensive experiments were used to demonstrate its security, inference accuracy comparable to plaintext approaches, and lower cost of secure inference.

Resonant and Ultra-short-period Planet Systems Are at Opposite Ends of the Exoplanet Age Distribution

Exoplanet systems are thought to evolve on secular timescales over billions of years. This evolution is impossible to directly observe on human timescales in most individual systems. While the availability of accurate and precise age inferences for individual exoplanet host stars with ages τ in the interval 1 Gyr ≲ τ ≲ 10 Gyr would constrain this evolution, accurate and precise age inferences are difficult to obtain for isolated field dwarfs like the host stars of most exoplanets. The Galactic velocity dispersion of a thin-disk stellar population monotonically grows with time, and the relationship between age and velocity dispersion in a given Galactic location can be calibrated by a stellar population for which accurate and precise age inferences are possible. Using a sample of subgiants with precise age inferences, we calibrate the age–velocity dispersion relation in the Kepler field. Applying this relation to the Kepler field’s planet populations, we find that Kepler-discovered systems plausibly in second-order mean-motion resonances have 1 Gyr ≲ τ ≲ 2 Gyr. The same is true for systems plausibly in first-order mean-motion resonances, but only for systems likely affected by tidal dissipation inside their innermost planets. These observations suggest that many planetary systems diffuse away from initially resonant configurations on secular timescales. Our calibrated relation also indicates that ultra-short-period (USP) planet systems have typical ages in the interval 5 Gyr ≲ τ ≲ 6 Gyr. We propose that USP planets tidally migrated from initial periods in the range 1 day ≲ P ≲ 2 days to their observed locations at P < 1 day over billions of years and trillions of cycles of secular eccentricity excitation and inside-planet damping.

A protocol for assessing bias and robustness of social network metrics using GPS based radio-telemetry data

BackgroundSocial network analysis of animal societies allows scientists to test hypotheses about social evolution, behaviour, and dynamic processes. However, the accuracy of estimated metrics depends on data characteristics like sample proportion, sample size, and frequency. A protocol is needed to assess for bias and robustness of social network metrics estimated for the animal populations especially when a limited number of individuals are monitored.MethodsWe used GPS telemetry datasets of five ungulate species to combine known social network approaches with novel ones into a comprehensive five-step protocol. To quantify the bias and uncertainty in the network metrics obtained from a partial population, we presented novel statistical methods which are particularly suited for autocorrelated data, such as telemetry relocations. The protocol was validated using a sixth species, the fallow deer, with a known population size where ∼85%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\sim 85\\%$$\\end{document} of the individuals have been directly monitored.ResultsThrough the protocol, we demonstrated how pre-network data permutations allow researchers to assess non-random aspects of interactions within a population. The protocol assesses bias in global network metrics, obtains confidence intervals, and quantifies uncertainty of global and node-level network metrics based on the number of nodes in the network. We found that global network metrics like density remained robust even with a lowered sample size, while local network metrics like eigenvector centrality were unreliable for four of the species. The fallow deer network showed low uncertainty and bias even at lower sampling proportions, indicating the importance of a thoroughly sampled population while demonstrating the accuracy of our evaluation methods for smaller samples.ConclusionsThe protocol allows researchers to analyse GPS-based radio-telemetry or other data to determine the reliability of social network metrics. The estimates enable the statistical comparison of networks under different conditions, such as analysing daily and seasonal changes in the density of a network. The methods can also guide methodological decisions in animal social network research, such as sampling design and allow more accurate ecological inferences from the available data. The R package aniSNA enables researchers to implement this workflow on their dataset, generating reliable inferences and guiding methodological decisions.