Simple Probabilistic Model Research Articles

Re-sampled inheritance compact optimization

Compact optimization is an alternative paradigm in the field of metaheuristics requiring a modest use of memory to optimize a problem. As opposed to population-based algorithms, which conduct the search by employing a set of candidate solutions, compact algorithms use a probabilistic model to describe how solutions are distributed over the search space. Compared to other Estimation of Distribution Algorithms, peculiar features such as the use of simple probabilistic models, in which variables are treated independently, and the need for a minimal number of solutions to be sampled to perform the search, make these algorithms suitable for those applications plagued by memory limitations. Compact algorithms show good results on different kinds of optimization problems but often prematurely converge and perform poorly on non-separable. In this paper, we attempt to overcome these limitations by combining compact algorithms with a restart mechanism named Re-Sampled Inheritance (RI) whose purpose is to avoid premature convergence while also inheriting parts of the variables from the best solution found so far. To assess the effect of the RI mechanism, we extensively test various existing compact algorithms, with and without RI, and compare the best RI-based compact algorithm against several competing algorithms on several optimization problems at different dimensionalities. We also evaluate the effect of the RI parameters on the overall algorithmic performance. Our numerical results not only show that RI consistently enhances the performances of compact algorithms, but also shed some light on the effectiveness of different compact logics at handling problems at different dimensionalities.

A Mathematical Model of Dignāga’s Hetu-cakra

A reasoned argument or tarka is essential for a wholesome vāda that aims at establishing the truth. A strong tarka constitutes of a number of elements including an anumāna based on a valid hetu. Several scholars, such as Dharmakīrti, Vasubandhu and Dignāga, have worked on theories for the establishment of a valid hetu to distinguish it from an invalid one. This paper aims to interpret Dignāga’s hetu-cakra, called the wheel of grounds, from a modern philosophical perspective by deconstructing it into a simple probabilistic mathematical model. The objective is to understand how and why a vāda based on a probabilistically weaker hetu can degrade into a Jalpa or vitaṇḍā. To do so, the paper maps the concept of ‘Bounded Rationality’ onto the hetu-cakra. Bounded Rationality, an idea coined by the management thinker Herbert Simon, is often employed in understanding decision-making processes of rational agents. In the context of this paper, the concept would state that the prativādin and ālocaka (debater) may not hold unbounded information to back their pratijnā (proposition). The paper argues that within the probabilistically deconstructed hetu-cakra model, most people argue in the ‘Zone of Bounded Rationality’, and thus, the probability of a debate degrading into Jalpa or vitaṇḍā is high.

The impact of community-wide, mass drug administration on aggregation of soil-transmitted helminth infection in human host populations

BackgroundSoil-transmitted helminths (STH) are intestinal parasites estimated to infect over 1.5 billion people. Current treatment programmes are aimed at morbidity control through school-based deworming programmes (targeting school-aged children, SAC) and treating women of reproductive age (WRA), as these two groups are believed to record the highest morbidity. More recently, however, the potential for interrupting transmission by treating entire communities has been receiving greater emphasis and the feasibility of such programmes are now under investigation in randomised clinical trials through the Bill & Melinda Gates Foundation funded DeWorm3 studies. Helminth parasites are known to be highly aggregated within human populations, with a small minority of individuals harbouring most worms. Empirical evidence from the TUMIKIA project in Kenya suggests that aggregation may increase significantly after anthelminthic treatment.MethodsA stochastic, age-structured, individual-based simulation model of parasite transmission is employed to better understand the factors that might induce this pattern. A simple probabilistic model based on compounded negative binomial distributions caused by age-dependencies in both treatment coverage and exposure to infection is also employed to further this understanding.ResultsBoth approaches confirm helminth aggregation is likely to increase post-mass drug administration as measured by a decrease in the value of the negative binomial aggregation parameter, k. Simple analytical models of distribution compounding describe the observed patterns well.ConclusionsThe helminth aggregation that was observed in the field was replicated with our stochastic individual-based model. Further work is required to generalise the probabilistic model to take account of the respective sensitivities of different diagnostics on the presence or absence of infection.

Sensory loss due to object formation

The precision to locate individual features in depth can often be improved by integrating information over space. However, this integration can sometimes be extremely detrimental, as for example in the case of the Westheimer-McKee phenomenon where features are grouped to form an object. We replicate here the known loss of precision in this phenomenon and document an additional loss of accuracy. These detrimental effects are still present when the object is elicited by other principles of organization, including a cross-modal auditory cue. Similar effects of object formation are found on lateral motion sensitivity. We then present a simple probabilistic model based on the integration of estimated depth within an object and propagation of object mean depth and uncertainty back to the elementary features of the object. This propagation of object uncertainty is a hitherto underestimated side-effect of object formation.

Global and regional probabilities of major nuclear reactor accidents

The continued and extended use of nuclear power is often considered and discussed as a viable energy policy option to meet energy demands while also meeting national CO2 emission reduction goals. A central issue in energy policy for sustainability is the question of nuclear reactor safety. However, studies on nuclear reactor safety often run up against the problem of estimating the probability of a major accident from patchy and limited empirical data. Here, we describe a simple probabilistic model of catastrophic nuclear reactor accidents based on a set of four assumptions. The model treats the accident probability in each of n reactors as a variable and returns the probability of a major accident in the reactor fleet. We find that, at 99.5% reactor safety, the probability of another Chernobyl- or Fukushima-sized event is 49% for the global fleet, and that safety would have to be 99.96% in order to bring that probability below 5%. We discuss our findings in light of the debate on energy policy for sustainability.

Object Extraction Using Probabilistic Maps of Color, Depth, and Near-Infrared Information

Object extraction is one of the important and chal-lenging tasks in the computer vision and/or robotics ? elds.This task is to extract the object from the scene using anypossible cues. The scenario discussed in this study was the objectextraction which considering the Space of Interest (SOI), i.e.,the three dimensional area where the object probably existed.To complete such task, the object extraction method based onthe probabilistic maps of multiple cues was proposed. Thanksto the Kinect V2 sensor, multiple cues such as color, depth, andnear-infrared information can be acquired simultaneously. TheSOI was modeled by a simple probabilistic model by consideringthe geometry of the possible objects and the reachability of thesystem acquired from depth information. To model the color andnear-infrared information, a Gaussian mixture models (GMM)was used. All of the models were combined to generate theprobabilistic maps that were used to extract the object fromthe scene. To validate the proposed object extraction, severalexperiments were conducted to investigate the best combinationof the cues used in this study.

Probabilistic model for time to cover cracking due to corrosion

AbstractThe time to cover cracking is widely used as a service life indicator in the assessment of deterioration of corrosion‐affected reinforced concrete structures. This paper presents a general probabilistic procedure for prediction of time to cover cracking. Within this procedure, the response surface method (RSM) is employed to calibrate a new model for calculating of radial displacement required for cover cracking based on the results obtained from the finite element (FE) analysis. By taking advantage of the central limit theorem (CLT), simple but accurate probabilistic models for prediction of time to cover cracking that only rely on the knowledge of first and second moments of basic random variables are derived. Rigorous simulation analysis has proved the accuracy of these models. It is shown in the paper that using this probabilistic procedure, factors affecting randomness of time to cover cracking can be easily identified. It is also shown that the time to cover cracking is highly variable with the concrete cover, tensile strength of concrete, corrosion current density, and the model error as the most influential factors on its randomness.

Foam stability in filtered lubricants containing antifoams

Lubricant formulations are filtered to remove deleterious particulate matter. An unintended consequence of this important process is the detrimental effect of fine filtration on the foaming performance of lubricants with antifoam additives. Here we outline a method to study this phenomenon in detail by probing the coalescence stability of single bubbles in filtered antifoam laden lubricants. Initially, we establish the validity of Garrett’s hypothesis for the tested antifoam laden lubricants. Subsequently, we show that the bubble stability in filtered lubricants are positively correlated to the number of filtration cycles - with the most dramatic changes in bubble stability accompanying the initial few cycles of filtration. Further, we show that post filtration, the stability of bubbles in lubricants is inversely correlated to the pore size of the filter and the volume fraction of antifoam in the lubricant prior to filtration. The results also reveal that in the presence of antifoam additives, the bubble coalescence times span multiple Rayleigh distributions. We also provide visual evidence that shows the tested antifoams employ a bridging-stretching mechanism to rupture non-aqueous foams. Finally, a simple probabilistic model is introduced that helps in analyzing the distribution of coalescence times of single bubbles to obtain insights into the volume fraction of antifoams in the lubricant. We believe these results are valuable in guiding the design of lubricants with robust and superior foaming performance.

Seismic reliability of axially loaded vertical piles

The reliability of single vertical pile foundations subjected to seismic loads is assessed and compared with the minimum acceptable reliability level for static load conditions mandated by the Canadian codes. The analysis is executed for a site with a mean shear-wave velocity of the top 30 m of the ground equal to 250 m/s subjected to the ground motion hazard of five Canadian cities. Using both a full probabilistic analysis and simplified probabilistic model, the results seem to indicate that the current design practice is unable to achieve the reliability target of the codes. The shortfall is particularly significant when the limiting pile settlement is relatively small. The calculated reliability level of small limiting settlements is impacted by the geotechnical variability, whereas the seismic hazard variability affects large pile limiting settlements. Finally, the simplified probabilistic model produces the same results as the full probabilistic model for large pile settlement and is a convenient tool to execute code calibration.

An Information-Theoretic Approach to Personalized Explainable Machine Learning

Automated decision making is used routinely throughout our everyday life. Recommender systems decide which jobs, movies, or other user profiles might be interesting to us. Spell checkers help us to make good use of language. Fraud detection systems decide if a credit card transactions should be verified more closely. Many of these decision making systems use machine learning methods that fit complex models to massive datasets. The successful deployment of machine learning (ML) methods to many (critical) application domains crucially depends on its explainability. Indeed, humans have a strong desire to get explanations that resolve the uncertainty about experienced phenomena like the predictions and decisions obtained from ML methods. Explainable ML is challenging since explanations must be tailored (personalized) to individual users with varying backgrounds. Some users might have received university-level education in ML, while other users might have no formal training in linear algebra. Linear regression with few features might be perfectly interpretable for the first group but might be considered a black-box by the latter. We propose a simple probabilistic model for the predictions and user knowledge. This model allows to study explainable ML using information theory. Explaining is here considered as the task of reducing the "surprise" incurred by a prediction. We quantify the effect of an explanation by the conditional mutual information between the explanation and prediction, given the user background.

Free will as a higher-level phenomenon?

List (2014, 2019) has recently argued for a particular view of free will as a higher-level phenomenon compatible with determinism. According to List, one could refute his account by showing that determinism at the physical level implies the impossibility of doing otherwise at the agential level. This paper takes up that challenge. Based on assumptions to which List's approach is committed, I provide a simple probabilistic model that establishes the connection between physical determinism and the impossibility of doing otherwise at the agential level that is needed to refute free will as a higher-level phenomenon.

A Conciliatory Answer to the Paradox of the Ravens

In the Paradox of the Ravens, a set of otherwise intuitive claims about evidence seems to be inconsistent. Most attempts at answering the paradox involve rejecting a member of the set, which seems to require a conflict either with commonsense intuitions or with some of our best confirmation theories. In contrast, I argue that the appearance of an inconsistency is misleading: ‘confirms’ and cognate terms feature a significant ambiguity when applied to universal generalisations. In particular, the claim that some evidence confirms a universal generalisation ordinarily suggests, in part, that the evidence confirms the reliability of predicting that something which satisfies the antecedent will also satisfy the consequent. I distinguish between the familiar relation of confirmation simpliciter and what I shall call ‘predictive confirmation’. I use them to formulate my answer, illustrate it in a very simple probabilistic model, and defend it against objections. I conclude that, once our evidential concepts are sufficiently clarified, there is no sense in which the initial claims are both plausible and inconsistent.

Self-Limited Accumulation of Colloids in Porous Media.

We present local direct imaging of the progressive adsorption of colloidal particles inside a 3D model porous medium. By varying the interparticle electrostatic interactions, we observe a large range of particle deposition regimes, from a single layer of particles at the surface of the medium to multiple layers and eventually clogging of the system. We derive the complete deposition dynamics and show that colloid accumulation is a self-limited mechanism towards a deposited fraction associated with a balance between the particle interactions and the imposed flow rate. These trends are explained and predicted using a simple probability model considering the particle adsorption energy and the variation of the drag energy with evolving porosity. This constitutes a direct validation of speculated particle transport mechanisms, and a further understanding of accumulation mechanisms.

Deep Trajectory Recovery with Fine-Grained Calibration using Kalman Filter

With the development of location-acquisition technologies, there are a huge number of mobile trajectories generated and accumulated in a variety of domains. However, due to the constraints of device and environment, many trajectories are recorded at low sampling rate, which increases the uncertainty between two consecutive sampled points in the trajectories. Our task is to recover a high-sampled trajectory based on the irregular low-sampled trajectory in free space, i.e., without road network information. There are two major problems with traditional solutions. First, many of these methods rely on heuristic search algorithms or simple probabilistic models. They cannot well capture complex sequential dependencies or global data correlations. Second, for reducing the predictive complexity of the unconstrained numerical coordinates, most of the previous studies have adopted a common preprocessing strategy by mapping the space into discrete units. As a side effect, using discrete units is likely to bring noise or inaccurate information. Hence, a principled post-calibration step is required to produce accurate results, which has been seldom studied by existing methods. To address the above difficulties, we propose a novel Deep Hybrid Trajectory Recovery model, named DHTR . Our recovery model extends the classic sequence-to-sequence generation framework by implementing a subsequence-to-sequence recovery model tailored for the current task, named subseq2seq. In order to effectively capture spatiotemporal correlations, we adopt both spatial and temporal attentions for enhancing the model performance. With the attention mechanisms, our model is able to characterize long-range correlations among trajectory points. Furthermore, we integrate the subseq2seq with a calibration component of Kalman filter (KF) for reducing the predictive uncertainty. At each timestep, the noisy predictions from the subseq2seq component will be fed into the KF component for calibration, and then the refined predictions will be forwarded to the subseq2seq component for the computation of the next timestep. Extensive results on real-world datasets have shown the superiority of the proposed model in both performance and interpretability.

Diabetes diagnosis from administrative claims and estimation of the true prevalence of diabetes among 4.2 million individuals of the Veneto region (North East Italy)

Background and aimsDiabetes can often remain undiagnosed or unregistered in administrative databases long after its onset, even when laboratory test results meet diagnostic criteria. In the present work, we analyse healthcare data of the Veneto Region, North East Italy, with the aims of: (i) developing an algorithm for the identification of diabetes from administrative claims (4,236,007 citizens), (ii) assessing its reliability by comparing its performance with the gold standard clinical diagnosis from a clinical database (7525 patients), (iii) combining the algorithm and the laboratory data of the regional Health Information Exchange (rHIE) system (543,520 subjects) to identify undiagnosed diabetes, and (iv) providing a credible estimate of the true prevalence of diabetes in Veneto. Methods and resultsThe proposed algorithm for the identification of diabetes was fed by administrative data related to drug dispensations, outpatient visits, and hospitalisations. Evaluated against a clinical database, the algorithm achieved 95.7% sensitivity, 87.9% specificity, and 97.6% precision. To identify possible cases of undiagnosed diabetes, we applied standard diagnostic criteria to the laboratory test results of the subjects who, according to the algorithm, had no diabetes-related claims. Using a simplified probabilistic model, we corrected our claims-based estimate of known diabetes (6.17% prevalence; 261,303 cases) to account for undiagnosed cases, yielding an estimated total prevalence of 7.50%. ConclusionWe herein validated an algorithm for the diagnosis of diabetes using administrative claims against the clinical diagnosis. Together with rHIE laboratory data, this allowed to identify possibly undiagnosed diabetes and estimate the true prevalence of diabetes in Veneto.

Sensitivity to Stimulus Irregularity Is Inherent in Neural Networks

Behavior is controlled by complex neural networks in which neurons process thousands of inputs. However, even short spike trains evoked in a single cortical neuron were demonstrated to be sufficient to influence behavior in vivo. Specifically, irregular sequences of interspike intervals (ISIs) had a more reliable influence on behavior despite their resemblance to stochastic activity. Similarly, irregular tactile stimulation led to higher rates of behavioral responses. In this study, we identify the mechanisms enabling this sensitivity to stimulus irregularity (SSI) on the neuronal and network levels using simulated spiking neural networks. Matching in vivo experiments, we find that irregular stimulation elicits more detectable network events (bursts) than regular stimulation. Dissecting the stimuli, we identify short ISIs-occurring more frequently in irregular stimulations-as the main drivers of SSI rather than complex irregularity per se. In addition, we find that short-term plasticity modulates SSI. We subsequently eliminate the different mechanisms in turn to assess their role in generating SSI. Removing inhibitory interneurons, we find that SSI is retained, suggesting that SSI is not dependent on inhibition. Removing recurrency, we find that SSI is retained due to the ability of individual neurons to integrate activity over short timescales ("cell memory"). Removing single-neuron dynamics, we find that SSI is retained based on the short-term retention of activity within the recurrent network structure ("network memory"). Finally, using a further simplified probabilistic model, we find that local network structure is not required for SSI. Hence, SSI is identified as a general property that we hypothesize to be ubiquitous in neural networks with different structures and biophysical properties. Irregular sequences contain shorter ISIs, which are the main drivers underlying SSI. The experimentally observed SSI should thus generalize to other systems, suggesting a functional role for irregular activity in cortex.

Towards crack paths simulations in media with a random fracture energy

Simple probabilistic models of fracture of 3D polycrystals involving transgranular and intergranular cracks are proposed, based on undamaged paths and appropriate percolation thresholds. In a second part, we propose theoretical extensions of phase field models for crack initiation and propagation to the case of locally heterogeneous and anisotropic fracture energy, and their implementation to get full field solutions by means of iterations of Fourier transforms to replace the standard finite element approach.

A fire safety assessment approach for evacuation analysis in underground physics research facilities

This paper introduces an integrated approach for evacuation assessment of nuclear physics research facilities exposed to fire risk. The approach combines the use of a simplified egress modelling method and advanced agent-based simulations of evacuation. An integrated multi-model approach is proposed here given the varying level of complexity concerning evacuation safety in underground physics facilities. This paper introduces a simplified probabilistic egress model based on existing hand calculations for 1D smoke spread modelling and it suggests a procedure for its combined use with advanced agent-based evacuation simulations. This includes the use of the outputs of the simplified egress model in underground smoke-filled portions of underground nuclear research facilities (e.g. tunnel arcs) as an input for complex agent-based evacuation simulations in the underground access shafts. An exemplary application of the integrated approach is presented for the simulation of a set of hypothetical fire risk scenarios in the Future Circular Collider (FCC) at CERN. This approach is deemed to facilitate fire evacuation safety assessment in underground physics research facilities by optimizing the simulation of relevant fire risk scenarios. A discussion on the advantages and implications of the use of an integrated approach in comparison with other safety assessment methods is presented.

Use of Bayesian models to develop coral bleaching indices and forecasts from in-situ observations for the 2015–16 bleaching event

Abstract Bayesian models provide simple conceptual probabilistic models of systems and system processes. Using real time data from IMOS Wireless Sensor Networks at four sites along the Great Barrier Reef, two Bayesian models were built to provide daily indices of current and future bleaching risk for the 2015–16 coral bleaching event. The first model used real time measures of water temperature and light, calculated degree-heat-week values and empirical bleaching thresholds to model the heat and light stress to produce a current bleaching risk index. The second model incorporated factors known to change bleaching risk, such as winds, tides and rainfall, to model the likely change in bleaching risk for the near future. The models used daily values from four IMOS Wireless Sensor Network sites, spread from north to south, for the 2015–16 bleaching event with the outputs delivered to the coral reef community via a dedicated web site. The model outputs were subsequently compared to aerial bleaching surveys with the models providing a good measure of the bleaching risk and, importantly, giving 2–3 days warning of extreme events such as the bleaching of northern reefs. These warnings were used to drive observational work to capture data from the bleaching event, an event that impacted a third of the Great Barrier Reef. Bayesian models provide easy to understand process or system models that can be quickly developed using a range of available inputs, including anecdotal and low-quality sources, to produce management relevant indices.

A simplified probabilistic model for the combined action of carbonation and chloride ingress

This paper presents an improved numerical model for the reliability analysis of corrosion initiation due to the combined action of carbonation and chloride ingress. The coupled mechanism of carbon dioxide, chloride ions, heat and moisture is considered in the model. The probability of corrosion initiation is evaluated by means of a single limit state function in which the critical concentration of chloride ions depends on the degree of carbonation. In addition, the coupled diffusion coefficients are interpolated linearly depending on the degree of carbonation in order to simplify the structure of the model without compromising accuracy.