Infinite Mixture Research Articles

New mitochondrial genomes of three whip spider species from the Amazon (Arachnida, Amblypygi) with phylogenetic relationships and comparative analysis

The complete mitochondrial genomes of the whip spiders Charinus carajas, C. ferreus, and Heterophrynus longicornis were sequenced, annotated, and compared with other mitogenomes of whip spiders and arachnids. The three new mitogenomes have the 37 genes usually observed in Metazoa: 13 protein-coding genes (PCGs), 22 transfer RNAs (tRNAs), and two ribosomal RNAs (rRNAs), plus a non-coding control region (CR). Most PCGs presented an ATN start codon, except cox1 in both Charinus species, initiating with TTA. Most PCGs terminated with stop codons TAA or TAG, except nad5 of C. carajas and cox3 of H. longicornis, which presented an incomplete stop codon (T). The Ka/Ks ratios were less than one for all the PCGs, indicating these genes are under purifying selection. All the tRNAs, except for serine 1 (trnS1), had the typical cloverleaf-shaped secondary structure. All the phylogenetic analyses resolved Charinus carajas and C. ferreus as monophyletic groups. Nonetheless, we did not recover the monophyly of Heterophrynus longicornis. The phylogenies under partitioned models did not recover suprageneric taxonomic groups as clades, but the Bayesian inference under the CAT infinite mixture model recovered the family Phrynidae and the superfamily Phrynoidea as monophyletic groups.

Modelling the heterogeneity of tourist spending in a mature destination: An approach through infinite mixture

Identifying tourists' preferences is essential for stakeholders to provide better products and services. Among the tools to classify such choices, expenditure segmentation is valuable for separating tourist groups with shared interests. The underlying idea of the (infinite) mixture model is that tourists spend on a specific activity depending on their preferences. However, the propensity to consume may be based on or influenced by the group to which the tourist belongs. Thus, such a tendency could increase depending on homogeneity or heterogeneity. This paper uses a compound distribution mixture to model the expenditure heterogeneity. The resulting mixture model derives from a multivariate Pareto (Lomax) distribution that is easy to implement and includes zero value in its support since it is empirically proven that a tourist's expenditure on some activity can be zero. Results show that once the spending on transport has been carried out, tourists prefer to spend more on food than other activities. Conditioned to the expense carried out on food, the mean expenditure on leisure activities is more significant than on transport. Finally, tourists would prefer to spend more on food than on transportation once they decide to spend on other activities.

Greening Separation and Purification of Proteins and Peptides.

The increasing awareness of environmental issues and the transition to green analytical chemistry (GAC) have gained popularity among academia and industry in recent years. One of the principles of GAC is the reduction and replacement of toxic solvents with more sustainable and environmentally friendly ones. This review gives an overview of the advances in applying green solvents as an alternative to the traditional organic solvents for peptide and protein purification and analysis by liquid chromatography (LC) and capillary electrophoresis (CE) methods. The feasibility of using greener LC and CE methods is demonstrated through several application examples; however, there is still plenty of room for new developments to fully realize their potential and to address existing challenges. Thanks to the tunable properties of designer solvents, such as ionic liquids and deep eutectic solvents, and almost infinite possible mixtures of components for their production, it is possible that some new designer solvents could potentially surpass the traditional harmful solvents in the future. Therefore, future research should focus mainly on developing new solvent combinations and enhancing analytical instruments to be able to effectively work with green solvents.

Dynamics-Adaptive Continual Reinforcement Learning via Progressive Contextualization.

A key challenge of continual reinforcement learning (CRL) in dynamic environments is to promptly adapt the reinforcement learning (RL) agent's behavior as the environment changes over its lifetime while minimizing the catastrophic forgetting of the learned information. To address this challenge, in this article, we propose DaCoRL, that is, dynamics-adaptive continual RL. DaCoRL learns a context-conditioned policy using progressive contextualization, which incrementally clusters a stream of stationary tasks in the dynamic environment into a series of contexts and opts for an expandable multihead neural network to approximate the policy. Specifically, we define a set of tasks with similar dynamics as an environmental context and formalize context inference as a procedure of online Bayesian infinite Gaussian mixture clustering on environment features, resorting to online Bayesian inference to infer the posterior distribution over contexts. Under the assumption of a Chinese restaurant process (CRP) prior, this technique can accurately classify the current task as a previously seen context or instantiate a new context as needed without relying on any external indicator to signal environmental changes in advance. Furthermore, we employ an expandable multihead neural network whose output layer is synchronously expanded with the newly instantiated context and a knowledge distillation regularization term for retaining the performance on learned tasks. As a general framework that can be coupled with various deep RL algorithms, DaCoRL features consistent superiority over existing methods in terms of stability, overall performance, and generalization ability, as verified by extensive experiments on several robot navigation and MuJoCo locomotion tasks.

Context-Based Meta-Reinforcement Learning With Bayesian Nonparametric Models.

Deep reinforcement learning agents usually need to collect a large number of interactions to solve a single task. In contrast, meta-reinforcement learning (meta-RL) aims to quickly adapt to new tasks using a small amount of experience by leveraging the knowledge from training on a set of similar tasks. State-of-the-art context-based meta-RL algorithms use the context to encode the task information and train a policy conditioned on the inferred latent task encoding. However, most recent works are limited to parametric tasks, where a handful of variables control the full variation in the task distribution, and also failed to work in non-stationary environments due to the few-shot adaptation setting. To address those limitations, we propose MEta-reinforcement Learning with Task Self-discovery (MELTS), which adaptively learns qualitatively different nonparametric tasks and adapts to new tasks in a zero-shot manner. We introduce a novel deep clustering framework (DPMM-VAE) based on an infinite mixture of Gaussians, which combines the Dirichlet process mixture model (DPMM) and the variational autoencoder (VAE), to simultaneously learn task representations and cluster the tasks in a self-adaptive way. Integrating DPMM-VAE into MELTS enables it to adaptively discover the multi-modal structure of the nonparametric task distribution, which previous methods using isotropic Gaussian random variables cannot model. In addition, we propose a zero-shot adaptation mechanism and a recurrence-based context encoding strategy to improve the data efficiency and make our algorithm applicable in non-stationary environments. On various continuous control tasks with both parametric and nonparametric variations, our algorithm produces a more structured and self-adaptive task latent space and also achieves superior sample efficiency and asymptotic performance compared with state-of-the-art meta-RL algorithms.

Unsupervised outlier detection using random subspace and subsampling ensembles of Dirichlet process mixtures

Probabilistic mixture models are recognized as effective tools for unsupervised outlier detection owing to their interpretability and global characteristics. Among these, Dirichlet process mixture models stand out as a strong alternative to conventional finite mixture models for both clustering and outlier detection tasks. Unlike finite mixture models, Dirichlet process mixtures are infinite mixture models that automatically determine the number of mixture components based on the data. Despite their advantages, the adoption of Dirichlet process mixture models for unsupervised outlier detection has been limited by challenges related to computational inefficiency and sensitivity to outliers in the construction of outlier detectors. Additionally, Dirichlet process Gaussian mixtures struggle to effectively model non-Gaussian data with discrete or binary features. To address these challenges, we propose a novel outlier detection method that utilizes ensembles of Dirichlet process Gaussian mixtures. This unsupervised algorithm employs random subspace and subsampling ensembles to ensure efficient computation and improve the robustness of the outlier detector. The ensemble approach further improves the suitability of the proposed method for detecting outliers in non-Gaussian data. Furthermore, our method uses variational inference for Dirichlet process mixtures, which ensures both efficient and rapid computation. Empirical analyses using benchmark datasets demonstrate that our method outperforms existing approaches in unsupervised outlier detection.

A Stochastic Recurrent Encoder Decoder Network for Multistep Probabilistic Wind Power Predictions.

In this article, a stochastic recurrent encoder decoder neural network (SREDNN), which considers latent random variables in its recurrent structures, is developed for the first time for the generative multistep probabilistic wind power predictions (MPWPPs). The SREDNN enables the stochastic recurrent model under the encoder-decoder framework to engage exogenous covariates to produce better MPWPP. The SREDNN consists of five components, the prior network, the inference network, the generative network, the encoder recurrent network, and the decoder recurrent network. The SREDNN is equipped with two critical advantages compared with conventional RNN-based methods. First, the integration over the latent random variable builds an infinite Gaussian mixture model (IGMM) as the observation model, which drastically increases the expressiveness of the wind power distribution. Secondly, hidden states of the SREDNN are updated in a stochastic way, which builds an infinite mixture of the IGMM for describing the ultimate wind power distribution and enables the SREDNN to model complex patterns across wind speed and wind power sequences. Computational experiments are conducted on a dataset of a commercial wind farm having 25 wind turbines (WTs) and two publicly assessable WT datasets to verify the advantages and effectiveness of the SREDNN for MPWPP. Experimental results show that the SREDNN achieves a lower negative form of the continuously ranked probability score (CRPS*) as well as a superior sharpness and comparable reliability of prediction intervals by comparing against considered benchmarking models. Results also show the clear benefit gained from considering latent random variables in SREDNN.

Understanding shape selectivity effects of hydroisomerization using a reaction equilibrium model.

We study important aspects of shape selectivity effects of zeolites for hydroisomerization of linear alkanes, which produces a myriad of isomers, particularly for long chain hydrocarbons. To investigate the conditions for achieving an optimal yield of branched hydrocarbons, it is important to understand the role of chemical equilibrium in these reversible reactions. We conduct an extensive analysis of shape selectivity effects of different zeolites for the hydroisomerization of C7 and C8 isomers at chemical reaction equilibrium conditions. The reaction ensemble Monte Carlo method, coupled with grand-canonical Monte Carlo simulations, is commonly used for computing reaction equilibrium of heterogeneous reactions. The computational demands become prohibitive for a large number of reactions. We used a faster alternative in which reaction equilibrium is obtained by imposing chemical equilibrium in the gas phase and phase equilibrium between the gas phase components and the adsorbed phase counterparts. This effectively mimics the chemical equilibrium distribution in the adsorbed phase. Using Henry's law at infinite dilution and mixture adsorption isotherm models at elevated pressures, we calculate the adsorbed loadings in the zeolites. This study shows that zeolites with cage or channel-like structures exhibit significant differences in selectivity for alkane isomers. We also observe a minimal impact of pressure on the gas-phase equilibrium of these reactions at typical experimental reaction temperatures 400-700K. This study marks initial strides in understanding the reaction product distribution for long-chain alkanes.

Early event detection for facial expression based on infinite mixture prototypes

In early event detection, the ability to detect events at an earlier stage means to have more ample time for processing. The purpose of this study is to apply few-shot learning to early event detection based on infinite mixture prototypes. we present a design to distribute the early event detection issues across the loss function and the data task setting in few-shot learning, and the adoption of infinite mixture models helps to resolve the multi-model distribution of the pattern of early events. The proposed algorithm is evaluated on the CK＋database. The experimental results demonstrate not only feasible performance in terms of accuracy and detection time but also alignment with our hypothesis through the observation of generated clusters.

Using fusion effects to decrease uncertainty in distance sampling models when collating data from different surveys

AbstractEstimates of population abundance are required to study the impacts of human activities on populations and assess their conservation status. Despite considerable effort to improve data collection, uncertainty around estimates of cetacean densities can remain large. A fundamental concept underlying distance sampling is the detection function. Here we focus on reducing the uncertainty in the estimation of detection function parameters in analyses combining data sets from multiple surveys, with known effects on the precision of density estimates. We developed detection functions using infinite mixture models that can be applied on data collating multiple species and/or surveys. These models enable automatic clustering by fusing the species and surveys with similar detection functions. We present a simulation analysis of a multisurvey data set in a Bayesian framework where we demonstrated that distance sampling models including fusion effects showed lower uncertainty than classical distance sampling models. We illustrated the benefits of this new model using data of line transect surveys from the Bay of Biscay and Iberian Coast. Future estimates of abundance using conventional distance sampling models on large multispecies surveys or on data sets combining multiple surveys could benefit from this new model to provide more precise density estimates.

Ant backbone phylogeny resolved by modelling compositional heterogeneity among sites in genomic data

Ants are the most ubiquitous and ecologically dominant arthropods on Earth, and understanding their phylogeny is crucial for deciphering their character evolution, species diversification, and biogeography. Although recent genomic data have shown promise in clarifying intrafamilial relationships across the tree of ants, inconsistencies between molecular datasets have also emerged. Here I re-examine the most comprehensive published Sanger-sequencing and genome-scale datasets of ants using model comparison methods that model among-site compositional heterogeneity to understand the sources of conflict in phylogenetic studies. My results under the best-fitting model, selected on the basis of Bayesian cross-validation and posterior predictive model checking, identify contentious nodes in ant phylogeny whose resolution is modelling-dependent. I show that the Bayesian infinite mixture CAT model outperforms empirical finite mixture models (C20, C40 and C60) and that, under the best-fitting CAT-GTR + G4 model, the enigmatic Martialis heureka is sister to all ants except Leptanillinae, rejecting the more popular hypothesis supported under worse-fitting models, that place it as sister to Leptanillinae. These analyses resolve a lasting controversy in ant phylogeny and highlight the significance of model comparison and adequate modelling of among-site compositional heterogeneity in reconstructing the deep phylogeny of insects.

Isothermal vapor-liquid equilibrium measurements and PC-SAFT, PR78, and CPA phase behavior modeling of n-tricosane + SC CO2 mixtures

This study reports the isothermal VLE (PTxy) data of n-tricosane (heavy n-alkane, n-C23H48) in supercritical (SC) CO2 at two selected temperatures of (323 and 343) K over the pressure range from (1.66 to 34.88) MPa. The measurements were made using a high-temperature and high-pressure VLE optical cell. The combined expanded uncertainty of the temperature, pressure, and phase concentration measurements at 0.95 confidence level with a coverage factor of k = 2 is estimated to be 0.2 K, 0.22 %, and 3.5 %, respectively. The measured VLE data has been used to estimate the critical parameters (TC, PC, xC) of the binary n-tricosane + CO2 mixture. The derived L-G critical properties data of the binary n-tricosane + CO2 mixture are: TC = 323.15 K, PC = 34.02 MPa, xC=0.060 mole fraction of n-tricosane; and TC = 343.15 K, PC = 40.14 MPa, xC=0.071 mole fraction of n-tricosane. The measured solubility data for the mixture has been used to estimate the physical meaning Krichevskii parameter ∂P∂xTCVC∞=-75.93MPa. Based on the derived Krichevskii parameter the thermodynamic (partial molar volumes, V¯2∞,H¯2∞,C¯P2∞) and structural (cluster size,Nexc∞) properties of the infinite dilution mixture of n-tricosane + CO2 near the critical point (CP) of pure solvent (CO2) have been calculated. The measured vapor – liquid equilibrium data for n-tricosane + CO2 was modeled using three equations of state (PC-SAFT, PR78, and CPA). The best prediction option with PC-SAFT was obtained with kij=0.12, while kij=0 was the best using PR78 and CPA EoS.

Nonparametric Bayesian modeling on infinite mixture Student t copulas

This article introduces an infinite mixture Student t copula model using a nonparametric Bayesian approach. We establish a corresponding Markov chain Monte Carlo sampler for this model. In contrast to the normal mixture model, our proposed model is more suitable for data exhibiting tail dependence, which is frequently encountered in financial risk management. We evaluated the proposed algorithm through theoretical simulations and real data analysis. Parameter estimation results from the simulations demonstrate that our approach is competitive when compared to the standard maximum likelihood estimation method. The analysis of real financial data supports the validity of our approach and highlights the importance of applying a t copula in the presence of heavy tails.

Quantifying the uncertainty of partitions for infinite mixture models

Bayesian clustering models, such as Dirichlet process mixture models (DPMMs), are sophisticated flexible models. They induce a posterior distribution on the set of all partitions of a set of observations. Analysing this posterior distribution is of great interest, but it comes with several challenges. First of all, the number of partitions is overwhelmingly large even for moderate values of the number of observations. Consequently the sample space of the posterior distribution of the partitions is not explored well by MCMC samplers. Second, due to the complexity of representing the uncertainty of partitions, usually only maximum a posteriori estimates of the posterior distribution of partitions are provided and discussed in the literature. In this paper we propose a numerical and graphical method for quantifying the uncertainty of the clusters of a given partition of the data and we suggest how this tool can be used to learn about the partition uncertainty.

Phylogenomics of weevils revisited: data curation and modelling compositional heterogeneity.

Weevils represent one of the most prolific radiations of beetles and the most diverse group of herbivores on land. The phylogeny of weevils (Curculionoidea) has received extensive attention, and a largely satisfactory framework for their interfamilial relationships has been established. However, a recent phylogenomic study of Curculionoidea based on anchored hybrid enrichment (AHE) data yielded an abnormal placement for the family Belidae (strongly supported as sister to Nemonychidae + Anthribidae). Here we reanalyse the genome-scale AHE data for Curculionoidea using various models of molecular evolution and data filtering methods to mitigate anticipated systematic errors and reduce compositional heterogeneity. When analysed with the infinite mixture model CAT-GTR or using appropriately filtered datasets, Belidae are always recovered as sister to the clade (Attelabidae, (Caridae, (Brentidae, Curculionidae))), which is congruent with studies based on morphology and other sources of molecular data. Although the relationships of the 'higher Curculionidae' remain challenging to resolve, we provide a consistent and robust backbone phylogeny of weevils. Our extensive analyses emphasize the significance of data curation and modelling across-site compositional heterogeneity in phylogenomic studies.

Experimental study and modeling of the isothermal VLE properties of ethylbenzene in supercritical solvents (CO2 and C3H8)

The isothermal VLE properties (PTxy) of two binary mixtures of ethylbenzene+CO2 and ethylbenzene+C3H8 at temperatures of (313.15, 333.15, 353.15) K, (393.15, 413.15, 433.15) K and at pressures up to 13 MPa and 6.4 MPa, respectively, were measured in the present work using a high-temperature and high-pressure optical cell VLE apparatus based on the static-analytic method. The combined expanded absolute uncertainty of the temperature, and the relative uncertainty of pressure, and concentration measurements at 0.95 confidence level with a coverage factor of k = 2 is estimated to be 0.15 K, 0.002 (or within 0.002–0.02 MPa, depending on temperature and pressure ranges), and 0.03, respectively, and 0.035 (or within 0.0001–0.028 mol fraction depending on temperature and pressure ranges), respectively. The critical parameters (TC, PC, xC) of the both mixtures were determined based on the measured isothermal phase equilibrium VLE data. Using the derived and reported critical property data (initial slopes of the critical curves at the solvent critical point, CO2 and C3H8) for ethylbenzene+CO2 and ethylbenzene+C3H8 mixtures, the values of the physical meaning (theoretically important) Krichevskii parameter have been calculated. The effect of the critical solvent (CO2 and C3H8) on the value of the Krichevskii parameter is studied. Pure- and mixture- like thermodynamic behavior (Fisher’s renormalization of the critical behavior of dilute mixtures) and structural properties (cluster sizes) of the infinite dilute mixture near the critical point of pure solvents (CO2 and C3H8) have been studied based on the Krichevskii parameter and initial slopes of the critical curves. Accuracies of the Perturbed-Chain Statistical Association Fluid Theory (PC-SAFT) and its Critical Point-based revision (CP-PC-SAFT) models in estimating the measured VLE data with the zero value of the binary adjustable parameter k12 were compared with E-PPR78 (Enhanced Predictive Peng–Robinson, 1978) approach.

Distributed online expectation-maximization algorithm for Poisson mixture model

Traffic flow data, in the form of multiple time series of aggregated traffic volume observed from various vehicle detector stations, are investigated as a motivating example. By the very nature of the traffic volume observations, we model the traffic volume data from each station by a Poisson mixture model. We then propose a distributed online expectation-maximization algorithm to fit the traffic volume data under the Poisson mixture model, with the results to be used to make inference about the population attributes of the traffic flows. Apart from the finite mixture Poisson model, two forms of infinite mixture Poisson – Poisson-inverse-Gaussian model and Poisson-Gamma model – are specifically considered and analyzed in detail. Simulation study and real data analysis are also undertaken to confirm the theoretical properties that we have derived for the proposed distributed online expectation-maximization algorithms.

Meta-learning representations for clustering with infinite Gaussian mixture models

Appropriate representations are critical for a better clustering performance. Although many neural network-based clustering methods have been proposed, they do not directly train neural networks to improve the clustering performance. We propose a method that can meta-learn knowledge for clustering from various labeled data and uses the knowledge for clustering unseen unlabeled data. The proposed method trains neural networks to obtain representations such that the clustering performance improves when the representations are clustered by variational Bayesian (VB) inference with an infinite Gaussian mixture model. For the objective function, we propose a continuous approximation of the adjusted Rand index (ARI) by which we can evaluate the clustering performance from soft clustering assignments. Since the approximated ARI and the VB inference procedure are differentiable, we can backpropagate the objective function through the VB inference procedure to train the neural networks. With experiments using text and image datasets, we demonstrate that our proposed method has a higher adjusted Rand index than the existing methods.

Fisher information and its extensions based on infinite mixture density functions

In this work, we consider the Fisher information and some of its well-known extended versions and then establish some results based on infinite mixture density functions for the proposed information measures. Specifically we introduce the Jensen-type of Fisher information (parametric-type and density-based) and generalized Fisher information measures based on infinite mixture density functions. We then show that the proposed Jensen–Fisher information measure have two representations in terms of Fisher information distance in the both parametric-type and density-based cases. We have further shown that the generalized Fisher information of an infinite mixture density can be stated based on Pearson–Vajda χk divergence. Finally, we examine the convexity property of q−Fisher information measure based on finite mixture density functions under a mild condition. Some examples related to scale mixture of skew-normal family of distributions, with emphasis on skew-normal density, are illustrated within the paper.

Bayesian Nonparametric Classification for Incomplete Data With a High Missing Rate: an Application to Semiconductor Manufacturing Data

During the semiconductor manufacturing process, predicting the yield of the semiconductor is an important problem. Early detection of defective product production in the manufacturing process can save huge production cost. The data generated from the semiconductor manufacturing process have characteristics of non-normal distributions, random missing patterns and high missing rate, which complicate the prediction of the yield. We propose the Dirichlet Process -Naive Bayes model (DPNB) that can simultaneously impute missing values and address classification problems. Since the DPNB is based on the infinite Gaussian mixture model, it can estimate complex data distributions and make predictions for missing datasets with some missing patterns due to nice properties of the Gaussian distribution. The DPNB also performs well for high missing rates since it uses all information of observed components. Experiments on various real datasets including semiconductor manufacturing data show that the DPNB has better performance than state-of-the-art methods in terms of predicting missing values and target variables as percentage of missing values increases.