A transformation strategy for process partitioning in hierarchical concurrent process networks

This thesis is concerned with the statistical learning of probabilistic models for graph-structured data. It addresses both the theoretical aspects of network modelling--like the learning of appropriate representations for networks--and the practical difficulties in developing the algorithms to perform inference for the proposed models. The first part of the thesis addresses the problem of discrete-time dynamic network modeling. The objective is to learn the common structure and the underlying interaction dynamics among the entities involved in the observed temporal network. Two probabilistic modeling frameworks are developed. First, a Bayesian nonparametric framework is proposed to capture the static latent community structure and the evolving node-community memberships over time. More specifically, the hierarchical gamma process is utilized to capture the underlying intra-community and inter-community interactions. The appropriate number of latent communities can be automatically estimated via the inherent shrinkage mechanism of the hierarchical gamma process prior. The gamma Markov process are constructed to capture the evolving node-community relations. As the Bernoulli-Poisson link function is used to map the binary edges to the latent parameter space, the proposed method scales with the number of non-zero edges. Hence, the proposed method is particularly well-fitted to model large sparse networks. Moreover, a time-dependent hierarchical gamma process dynamic network model is proposed to capture the birth and death dynamics of the underlying communities. For performance evaluation, the proposed methods are compared with state-of-the-art statistical network models on both synthetic and real-world data. In the second part of the thesis, the main objective is to analyze continuous-time event-based dynamic networks. A fundamental problem in modeling such continuously-generated temporal interaction events data is to capture the reciprocal nature of the interactions among entities--the actions performed by one individual toward another increase the probability that an action of the same type to be returned. Hence, the mutually-exciting Hawkes process is utilized to capture the reciprocity between each pair of individuals involved in the observed dynamic network. In particular, the base rate of the Hawkes process is built upon the latent parameters inferred using the hierarchical gamma process edge partition model, to capture the underlying community structure. Moreover, each interaction event between two individuals is augmented with a pair of latent variables, which will be referred to as latent patterns, to indicate which of their involved communities lead to the occurring of that interaction. Accordingly, the proposed model allows the excitatory effects of each interaction on its opposite direction are determined by its latent patterns. Efficient Gibbs sampling and Expectation Maximization algorithms are developed to perform inference. Finally, the evaluations performed on the real-world data demonstrate the interpretability and competitive performance of the model compared with state-of-the-art methods. In the third part of this thesis, the objective is to analyze the common structure of multiple related data sources under the generative framework. First, a Bayesian nonparametric group factor analysis method is developed to factorize multiple related groups of data into the common latent factor space. The hierarchical beta Bernoulli process is exploited to induce sparsity over the group-specific factor loadings to strengthen the model interpretability. A collapsed variational inference scheme is proposed to perform efficient inference for large-scale data analysis in real-world applications. Moreover, a Poisson gamma memberships framework is investigated for joint modelling of network and related node features.

Collembola are a basal group of Hexapoda renowned for both unique morphological characters and significant ecological roles. However, a robust and plausible phylogenetic relationship between its deeply divergent lineages has yet to be achieved. We carried out a mitophylogenomic study based on a so far the most comprehensive mitochondrial genome dataset. Our data matrix contained mitogenomes of 31 species from almost all major families of all four orders, with 16 mitogenomes newly sequenced and annotated. We compared the linear arrangements of genes along mitochondria across species. Then we conducted 13 analyses each under a different combination of character coding, partitioning scheme and heterotachy models, and assessed their performance in phylogenetic inference. Several hypothetical tree topologies were also tested. Mitogenomic structure comparison revealed that most species share the same gene order of putative ancestral pancrustacean pattern, while seven species from Onychiuridae, Poduridae and Symphypleona bear different levels of gene rearrangements, indicating phylogenetic signals. Tomoceroidea was robustly recovered for the first time in the presence of all its families and subfamilies. Monophyly of Onychiuroidea was supported using unpartitioned models alleviating LBA. Paronellidae was revealed polyphyletic with two subfamilies inserted independently into Entomobryidae. Although Entomobryomorpha has not been well supported, more than half of the analyses obtained convincing topologies by placing Tomoceroidea within or near remaining Entomobryomorpha. The relationship between elongate-shaped and spherical-shaped collembolans still remained ambiguous, but Neelipleona tend to occupy the basal position in most trees. This study showed that mitochondrial genomes could provide important information for reconstructing the relationships among Collembola when suitable analytical approaches are implemented. Of all the data refining and model selecting schemes used in this study, the combination of nucleotide sequences, partitioning model and exclusion of third codon positions performed better in generating more reliable tree topology and higher node supports than others.

Partitioning is a commonly used method in phylogenetics that aims to accommodate variation in substitution patterns among sites. Despite its popularity, there have been few systematic studies of its effects on phylogenetic inference, and there have been no studies that compare the effects of different approaches to partitioning across many empirical data sets. In this study, we applied four commonly used approaches to partitioning to each of 34 empirical data sets, and then compared the resulting tree topologies, branch-lengths, and bootstrap support estimated using each approach. We find that the choice of partitioning scheme often affects tree topology, particularly when partitioning is omitted. Most notably, we find occasional instances where the use of a suboptimal partitioning scheme produces highly supported but incorrect nodes in the tree. Branch-lengths and bootstrap support are also affected by the choice of partitioning scheme, sometimes dramatically so. We discuss the reasons for these effects and make some suggestions for best practice.

BackgroundGenome level analyses have enhanced our view of phylogenetics in many areas of the tree of life. With the production of whole genome DNA sequences of hundreds of organisms and large-scale EST databases a large number of candidate genes for inclusion into phylogenetic analysis have become available. In this work, we exploit the burgeoning genomic data being generated for plant genomes to address one of the more important plant phylogenetic questions concerning the hierarchical relationships of the several major seed plant lineages (angiosperms, Cycadales, Gingkoales, Gnetales, and Coniferales), which continues to be a work in progress, despite numerous studies using single, few or several genes and morphology datasets. Although most recent studies support the notion that gymnosperms and angiosperms are monophyletic and sister groups, they differ on the topological arrangements within each major group.MethodologyWe exploited the EST database to construct a supermatrix of DNA sequences (over 1,200 concatenated orthologous gene partitions for 17 taxa) to examine non-flowering seed plant relationships. This analysis employed programs that offer rapid and robust orthology determination of novel, short sequences from plant ESTs based on reference seed plant genomes. Our phylogenetic analysis retrieved an unbiased (with respect to gene choice), well-resolved and highly supported phylogenetic hypothesis that was robust to various outgroup combinations.ConclusionsWe evaluated character support and the relative contribution of numerous variables (e.g. gene number, missing data, partitioning schemes, taxon sampling and outgroup choice) on tree topology, stability and support metrics. Our results indicate that while missing characters and order of addition of genes to an analysis do not influence branch support, inadequate taxon sampling and limited choice of outgroup(s) can lead to spurious inference of phylogeny when dealing with phylogenomic scale data sets. As expected, support and resolution increases significantly as more informative characters are added, until reaching a threshold, beyond which support metrics stabilize, and the effect of adding conflicting characters is minimized.

Mitogenome sequences stabilize the phylogenetics of weevils (Curculionoidea) and establish the monophyly of larval ectophagy

Evaluation of options for energy recovery from municipal solid waste in India using the hierarchical analytical network process

Network embedding aims at transferring node proximity in networks into distributed vectors, which can be leveraged in various downstream applications. Recent research has shown that nodes in a network can often be organized in latent hierarchical structures, but without a particular underlying taxonomy, the learned node embedding is less useful nor interpretable. In this work, we aim to improve network embedding by modeling the conditional node proximity in networks indicated by node labels residing in real taxonomies. In the meantime, we also aim to model the hierarchical label proximity in the given taxonomies, which is too coarse by solely looking at the hierarchical topologies. To this end, we propose TaxoGAN to co-embed network nodes and hierarchical labels, through a hierarchical network generation process. Particularly, TaxoGAN models the child labels and network nodes of each parent label in an individual embedding space while learning to transfer network proximity among the spaces of hierarchical labels through stacked network generators and embedding encoders. To enable robust and efficient model inference, we further develop a hierarchical adversarial training process. Comprehensive experiments and case studies on four real-world datasets of networks with hierarchical labels demonstrate the utility of TaxoGAN in improving network embedding on traditional tasks of node classification and link prediction, as well as novel tasks like conditional proximity search and fine-grained taxonomy layout.

Edge-guided and hierarchical aggregation network for robust medical image segmentation

Hierarchical delay-memory echo state network: A model designed for multi-step chaotic time series prediction

Vision plays an important role in the aesthetic cognition of human beings. When creating dance choreography, human dancers, who always observe their own dance poses in a mirror, understand the aesthetics of those poses and aim to improve their dancing performance. In order to develop artificial intelligence, a robot should establish a similar mechanism to imitate the above human dance behaviour. Inspired by this, this paper designs a way for a robot to visually perceive its own dance poses and constructs a novel dataset of dance poses based on real NAO robots. On this basis, this paper proposes a hierarchical processing network-based approach to automatic aesthetics evaluation of robotic dance poses. The hierarchical processing network first extracts the primary visual features by using three parallel CNNs, then uses a synthesis CNN to achieve high-level association and comprehensive processing on the basis of multi-modal feature fusion, and finally makes an automatic aesthetics decision. Notably, the design of this hierarchical processing network is inspired by the research findings in neuroaesthetics. Experimental results show that our approach can achieve a high correct ratio of aesthetic evaluation at 82.3%, which is superior to the existing methods.

Edge detection is one of the fundamental components of advanced computer vision tasks, and it is essential to preserve computational resources while ensuring a certain level of performance. In this paper, we propose a lightweight edge detection network called the Parallel and Hierarchical Network (PHNet), which draws inspiration from the parallel processing and hierarchical processing mechanisms of visual information in the visual cortex neurons and is implemented via a convolutional neural network (CNN). Specifically, we designed an encoding network with parallel and hierarchical processing based on the visual information transmission pathway of the "retina-LGN-V1" and meticulously modeled the receptive fields of the cells involved in the pathway. Empirical evaluation demonstrates that, despite a minimal parameter count of only 0.2 M, the proposed model achieves a remarkable ODS score of 0.781 on the BSDS500 dataset and ODS score of 0.863 on the MBDD dataset. These results underscore the efficacy of the proposed network in attaining superior edge detection performance at a low computational cost. Moreover, we believe that this study, which combines computational vision and biological vision, can provide new insights into edge detection model research.

Quality assessment of natural images is influenced by perceptual mechanisms, e.g., attention and contrast sensitivity, and quality perception can be generated in a hierarchical process. This paper proposes an architecture of Attention Integrated Hierarchical Image Quality networks (AIHIQnet) for no-reference quality assessment. AIHIQnet consists of three components: general backbone network, perceptually guided neck network, and head network. Multi-scale features extracted from the backbone network are fused to simulate image quality perception in a hierarchical manner. The attention and contrast sensitivity mechanisms modelled by an attention module capture essential information for quality perception. Considering that image rescaling potentially affects perceived quality, appropriate pooling methods in the non-convolution layers in AIHIQnet are employed to accept images with arbitrary resolutions. Comprehensive experiments on publicly available databases demonstrate outstanding performance of AIHIQnet compared to state-of-the-art models. Ablation experiments were performed to investigate the variants of the proposed architecture and reveal importance of individual components.

Molecular phylogeny of the aquatic beetle family Noteridae (Coleoptera: Adephaga) with an emphasis on data partitioning strategies

Morphological data are a fundamental source of evidence to reconstruct the Tree of Life, and Bayesian phylogenetic methods are increasingly being used for this task. Bayesian phylogenetic analyses require the use of evolutionary models, which have been intensively studied in the past few years, with significant improvements to our knowledge. Notwithstanding, a systematic evaluation of the performance of partitioned models for morphological data has never been performed. Here we evaluate the influence of partitioned models, defined by anatomical criteria, on the precision and accuracy of summary tree topologies considering the effects of model misspecification. We simulated datasets using partitioning schemes, trees, and other properties obtained from two empirical datasets, and conducted Bayesian phylogenetic analyses. Additionally, we reanalyzed 32 empirical datasets for different groups of vertebrates, applying unpartitioned and partitioned models, and, as a focused study case, we reanalyzed a dataset including living and fossil armadillos, testing alternative partitioning hypotheses based on functional and ontogenetic modules. We found that, in general, partitioning by anatomy has little influence on summary topologies analyzed under alternative partitioning schemes with a varying number of partitions. Nevertheless, models with unlinked branch lengths, which account for heterotachy across partitions, improve topological precision at the cost of reducing accuracy. In some instances, more complex partitioning schemes led to topological changes, as tested for armadillos, mostly associated with models with unlinked branch lengths. We compare our results with other empirical evaluations of morphological data and those from empirical and simulation studies of the partitioning of molecular data, considering the adequacy of anatomical partitioning relative to alternative methods of partitioning morphological datasets. [Evolutionary rates; heterogeneity; morphology; Mk model; partition; topology.].

The mode of E-3DMesh network with a large number of faulty nodes was investigated.Based on partitioning strategies,a new probabilistic analysis approach was given,which enabled to derive the node failure probability of E-3DMesh networks when the connectivity probability of E-3DMesh network was attributed.In order to remain connected with probability larger than 99% in E-3DMesh networks with millions of nodes,the network node failure probability was controlled below 3.86% was proved.The scheme is applicable to the study of other hierarchical network structures and of other network communication problems.