Hierarchical Constraints Research Articles

Fine-Grained Feature Generation for Generalized Zero-Shot Video Classification.

Generalized zero-shot video classification aims to train a classifier to classify videos including both seen and unseen classes. Since the unseen videos have no visual information during training, most existing methods rely on the generative adversarial networks to synthesize visual features for unseen classes through the class embedding of category names. However, most category names only describe the content of the video, ignoring other relational information. As a rich information carrier, videos include actions, performers, environments, etc., and the semantic description of the videos also express the events from different levels of actions. In order to use fully explore the video information, we propose a fine-grained feature generation model based on video category name and its corresponding description texts for generalized zero-shot video classification. To obtain comprehensive information, we first extract content information from coarse-grained semantic information (category names) and motion information from fine-grained semantic information (description texts) as the base for feature synthesis. Then, we subdivide motion into hierarchical constraints on the fine-grained correlation between event and action from the feature level. In addition, we propose a loss that can avoid the imbalance of positive and negative examples to constrain the consistency of features at each level. In order to prove the validity of our proposed framework, we perform extensive quantitative and qualitative evaluations on two challenging datasets: UCF101 and HMDB51, and obtain a positive gain for the task of generalized zero-shot video classification.

A fast non-convex optimization technique for human action recovery from misrepresented 3D motion capture data using trajectory movement and pair-wise hierarchical constraints

A fast non-convex optimization technique for human action recovery from misrepresented 3D motion capture data using trajectory movement and pair-wise hierarchical constraints

A Correlation Analysis Method for Geographical Object Flows from a Geoeconomic Perspective

Geographic object flow is the reason behind the relationship of geographic units. There are interactions in the process of dynamic change of a geographic object flow, and its regularity, which can reflect the relationship or pattern of geographic units in a region. In this paper, an association rule mining method for the geographic object flow linkage process is studied from a geoeconomics perspective. Additionally, an association rule mining algorithm with hierarchical constraints is proposed. Data segmentation is performed according to the time series characteristics of geographic object flow data. The basic attributes for the association rule mining are determined based on the basic parameters of geographic object flows, and a database for the association rule mining is formed according to the characteristics of the hierarchical structure of the geographic object flows. Based on the obtained data, the association rule mining algorithm with hierarchical constraints obtained using a parent–child matrix is improved by adding the Apriori algorithm. With the Indo-Pacific region as an example, the trade flow association rules for 25 countries in the region from 2010 to 2021 are selected. In addition, a mathematical statistical analysis of the strongly associated mined trade flows and geoeconomic factors is conducted in terms of both a basic feature analysis of trade flow associations and a country-oriented trade flow association analysis by considering domain knowledge. The effectiveness of the method has been evaluated from various perspectives such as correlation analysis, mathematical statistics, and comparison with the findings of existing studies and proved the validity of the method.

Priority relations and cooperation with multiple activity levels

This paper analyzes cooperation situations between heterogeneous players. It considers two types of heterogeneity. First, the players are differentiated with respect to a priority structure. This structure captures asymmetries between players, which may reflect exogenous rights, different needs, merit, or hierarchical constraints. Second, each player may have different cooperation possibilities represented by a set of activity levels. To analyze these situations, we enrich the model of multi-choice games, which is a natural extension of transferable utility games, with a priority structure. A new value on the class of multi-choice games with a priority structure is introduced. To accommodate the different activity levels and the asymmetries between players, this value follows an allocation process based on a lexicographic procedure. New axioms for multi-choice games with a priority structure are introduced. These axioms endogenously determine the lexicographic procedure used to define the value. Two axiomatic characterizations of this value are provided.

Exploring Low-Toxicity Chemical Space with Deep Learning for Molecular Generation.

Creating a wide range of new compounds that not only have ideal pharmacological properties but also easily pass long-term toxicity evaluation is still a challenging task in current drug discovery. In this study, we developed a conditional generative model by combining a semisupervised variational autoencoder (SSVAE) with an MGA toxicity predictor. Our aim is to generate molecules with low toxicity, good drug-like properties, and structural diversity. For multiobjective optimization, we have developed a method with hierarchical constraints on the toxicity space of small molecules to generate drug-like small molecules, which can also minimize the effect on the diversity of generated results. The evaluation results of the metrics indicate that the developed model has good effectiveness, novelty, and diversity. The generated molecules by this model are mainly distributed in low-toxicity regions, which suggests that our model can efficiently constrain the generation of toxic structures. In contrast to simply filtering toxic ones after generation, the low-toxicity molecular generative model can generate molecules with structural diversity. Our strategy can be used in target-based drug discovery to improve the quality of generated molecules with low-toxicity, drug-like, and highly active properties.

On Mann implicit composite subgradient extragradient methods for general systems of variational inequalities with hierarchical variational inequality constraints

In a real Hilbert space, let the VIP, GSVI, HVI, and CFPP denote a variational inequality problem, a general system of variational inequalities, a hierarchical variational inequality, and a common fixed-point problem of a countable family of uniformly Lipschitzian pseudocontractive mappings and an asymptotically nonexpansive mapping, respectively. We design two Mann implicit composite subgradient extragradient algorithms with line-search process for finding a common solution of the CFPP, GSVI, and VIP. The suggested algorithms are based on the Mann implicit iteration method, subgradient extragradient method with line-search process, and viscosity approximation method. Under mild assumptions, we prove the strong convergence of the suggested algorithms to a common solution of the CFPP, GSVI, and VIP, which solves a certain HVI defined on their common solutions set.

Hierarchical multi-label classification based on LSTM network and Bayesian decision theory for LncRNA function prediction

Growing evidence shows that long noncoding RNAs (lncRNAs) play an important role in cellular biological processes at multiple levels, such as gene imprinting, immune response, and genetic regulation, and are closely related to diseases because of their complex and precise control. However, most functions of lncRNAs remain undiscovered. Current computational methods for exploring lncRNA functions can avoid high-throughput experiments, but they usually focus on the construction of similarity networks and ignore the certain directed acyclic graph (DAG) formed by gene ontology annotations. In this paper, we view the function annotation work as a hierarchical multilabel classification problem and design a method HLSTMBD for classification with DAG-structured labels. With the help of a mathematical model based on Bayesian decision theory, the HLSTMBD algorithm is implemented with the long-short term memory network and a hierarchical constraint method DAGLabel. Compared with other state-of-the-art algorithms, the results on GOA-lncRNA datasets show that the proposed method can efficiently and accurately complete the label prediction work.

Hierarchical Modeling of Carbonate Fault-Controlled Paleokarst Systems: A Case Study of the Ordovician Reservoir in the Tahe Oilfield, Tarim Basin

As a special type of fracture-cavity carbonate reservoir, fault/fracture-controlled paleokarsts are affected by large strike-slip faults with various storage spaces, irregular geometry, and stochastic spatial distribution, which brings new challenges to their 3D quantitative characterization and geological modeling. Based on multi-scale research data, a hierarchical scheme of a fault-controlled paleokarst is established for the Tahe oilfield, and then a hierarchical modeling method for the paleokarst is proposed based on the principles of hierarchical constraint and genetic control. The results show that the fault-controlled paleokarst hierarchy is divided into four levels according to genesis and scale. These are the strike-slip fault impact zone, the fault-controlled external paleokarst envelope, the fault-controlled internal paleokarst architecture, and the internal karst cave fillings. The strike-slip fault impact zone model is established using deterministic data of the main fault, secondary fault, envelope range, and key geological horizons obtained from seismic interpretation. Based on the seismic FL (Fault Likelihood) attribute, the external envelope model of the fault-controlled paleokarst is established using the deterministic method through automatic attribute segmentation and drilling calibration. With the constraint of the external envelope, the internal architectural elements are classified and modeled. A large-scale karst cave model and mesoscale discrete fracture distribution model are established using the deterministic method based on the seismic texture property and seismic ant property truncation. Taking well data as conditional data and seismic-geological probability cubes as constraint data, the sequential indication simulation and object-based marked point process simulation methods are used to obtain the model of the dissolution pores distribution and the model of the discrete small-scale fractures distribution, respectively. Finally, a typical fault-controlled paleokarst reservoir in the TP area of the Tahe oilfield is taken as an example to test the above modeling method; the 3D integrated model developed in this study can reflect the spatial hierarchy of the fault-controlled paleokarst carbonate reservoir.

Cannabis tourists' perceived constraints to engaging in commercial cannabis tourism overseas: a comparison of first-time and repeat tourists

PurposeGiven the increasing number of tourists exposed to commercially available cannabis, it is important to understand visitors' perceived constraints to cannabis consumption while travelling. This study aims to compare cannabis tourists' perceived constraints between first-time and repeat tourists to gain comprehensive understanding.Design/methodology/approachThis study involved 32 conversational field interviews (5–10 min) with Chinese tourists who had smoked cannabis in coffee shops in Amsterdam to identify perceived constraints to engaging in cannabis tourism. The hierarchical constraint model (HCM) informed qualitative data coding using a deductive approach to compare first-time and repeat tourists. Content analysis was conducted manually.FindingsResults showed that both tourist groups faced intrapersonal, interpersonal and structural perceived constraints to cannabis consumption. Differences in first-time and repeat cannabis tourists' perceived constraints were explained using the neutralization technique framework and psychological tactics, such as the defence of necessity, claims of entitlement, normal practice and claims of relative acceptability. In addition, social exchange theory was employed to explain why repeat cannabis tourists perceived fewer constraints than first-time tourists.Practical implicationsAlthough the current study did not intend to address the importance of cannabis tourism in places where cannabis is commercially available, findings offer empirical guidance for industry practitioners and policymakers regarding cannabis use education, prevention and policy in tourism contexts; related efforts can promote the sustainable development of this tourism market while protecting cannabis tourists' physical and psychological well-being.Originality/valueThis study makes important theoretical contributions regarding travel constraints in the unique context of cannabis tourism between first-time and repeat tourists. Findings will also enable academics, industry operators, policymakers and local residents of cannabis tourism destinations to better understand how these tourists decide whether to consume cannabis overseas. Differences in these perceived constraints between first-time and repeat tourists are discussed to highlight the dynamic nature of travel constraints.

An intermediate interpolation of VR glasses display animation based on hierarchical constraints

An intermediate interpolation of VR glasses display animation based on hierarchical constraints

An intermediate interpolation of VR glasses display animation based on hierarchical constraints

An intermediate interpolation of VR glasses display animation based on hierarchical constraints

Financing Local Government and Development in Canada in the Aftermath of a Global Pandemic: Continuity and Change

In the post-pandemic environment, local governments must confront a challenging fiscal environment. Drawing on insights provided over the years by Richard Bird, the author re-examines certain pervasive themes found in discussions and analysis of municipal finance, such as the merits of benefits-based taxation at the local level, the hierarchical constraints on municipal fiscal decision making, and the reality of a perpetual fiscal crisis at the local level. The underlying issue of having to finance a broad set of expenditures on public services from a narrow revenue base has resulted in intergovernmental tensions and continuing debate over the capacity of the property tax base to meet the demands placed on it. Financing local government can be described in terms of borrowing, spending, and taxing. Each of these areas of municipal fiscal decision making was affected by the COVID-19 pandemic. Borrowing rules remained hierarchically constrained, pressure to increase expenditures grew, and tax and fee-based revenues were adversely affected. Further, it seems likely that the municipal fiscal base will be narrowed given the controversy around development charges. The author uses data from Ontario to illustrate the impact of the pandemic on municipal fiscal health. In addition, the paper includes the Richard Bird Urban Fiscal Health Dashboard to illustrate aspects of the long-run fiscal health of Ontario's local governments. Throughout the paper, the author raises questions that merit further research informed by the perspective that Richard Bird brought to our understanding of the mechanics and implications of local fiscal decision making.

Hierarchical Multilabel Ship Classification in Remote Sensing Images Using Label Relation Graphs

Hierarchical multilabel classification (HMC) assigns multiple labels to each instance with the labels organized under hierarchical relations. In ship classification in remote sensing images, depending on the expert knowledge and image quality, the same type of ships in different remote sensing images may be annotated with different class labels from coarse to fine levels such as merchant ship (MS) or container ship (CTS). In this article, we propose a novel deep network with two output channels and their associated loss functions to learn an HMC classifier using samples labeled at different levels in the hierarchy. In the proposed network, a hierarchy and exclusion (HEX) graph is introduced to model the label hierarchy, which satisfies hierarchical constraints by encoding semantic relations between any two labels. The output nodes of the first channel are organized according to the HEX graph, and its corresponding probabilistic classification loss is built to reflect the hierarchical structure of the HEX graph. On the other hand, the output nodes of the second channel only represent the finest grained (last level in the hierarchy) classes, and its multiclass cross-entropy loss is designed to enhance the discriminative power of the HMC classifier on the last level labels, which is also compatible with constraints in the HEX graph. The combination of these two losses from two output channels can effectively transfer the hierarchical information of ship taxonomy during network training. Experimental results on two commonly used ship datasets demonstrate that the proposed method outperforms the state-of-the-art HMC approaches, and is especially advantageous when trained with fewer fine-grained samples.

Hypothesis Testing for Hierarchical Structures in Cognitive Diagnosis Models

Cognitive Diagnosis Models (CDMs) are a special family of discrete latent variable models widely used in educational, psychological and social sciences. In many applications of CDMs, certain hierarchical structures among the latent attributes are assumed by researchers to characterize their dependence structure. Specifically, a directed acyclic graph is used to specify hierarchical constraints on the allowable configurations of the discrete latent attributes. In this paper, we consider the important yet unaddressed problem of testing the existence of latent hierarchical structures in CDMs. We first introduce the concept of testability of hierarchical structures in CDMs and present sufficient conditions. Then we study the asymptotic behaviors of the likelihood ratio test (LRT) statistic, which is widely used for testing nested models. Due to the irregularity of the problem, the asymptotic distribution of LRT becomes nonstandard and tends to provide unsatisfactory finite sample performance under practical conditions. We provide statistical insights on such failures, and propose to use parametric bootstrap to perform the testing. We also demonstrate the effectiveness and superiority of parametric bootstrap for testing the latent hierarchies over non-parametric bootstrap and the naïve Chi-squared test through comprehensive simulations and an educational assessment dataset.

Gaussian Mixture Model of Ground Filtering Based on Hierarchical Curvature Constraints for Airborne Lidar Point Clouds

This paper proposes a Gaussian mixture model of a ground filtering method based on hierarchical curvature constraints. Firstly, the thin plate spline function is iteratively applied to interpolate the reference surface. Secondly, gradually changing grid size and curvature threshold are used to construct hierarchical constraints. Finally, an adaptive height difference classifier based on the Gaussian mixture model is proposed. Using the latent variables obtained by the expectation-maximization algorithm, the posterior probability of each point is computed. As a result, ground and objects can be marked separately according to the calculated possibility. 15 data samples provided by the International Society for Photogrammetry and Remote Sensing are used to verify the proposed method, which is also compared with eight classical filtering algorithms. Experimental results demonstrate that the average total errors and average Cohen's kappa coefficient of the proposed method are 6.91% and 80.9%, respectively. In general, it has better performance in areas with terrain discontinuities and bridges.

Approximate Laplace Approximations for Scalable Model Selection

AbstractWe propose the approximate Laplace approximation (ALA) to evaluate integrated likelihoods, a bottleneck in Bayesian model selection. The Laplace approximation (LA) is a popular tool that speeds up such computation and equips strong model selection properties. However, when the sample size is large or one considers many models the cost of the required optimizations becomes impractical. ALA reduces the cost to that of solving a least-squares problem for each model. Further, it enables efficient computation across models such as sharing pre-computed sufficient statistics and certain operations in matrix decompositions. We prove that in generalized (possibly non-linear) models ALA achieves a strong form of model selection consistency for a suitably-defined optimal model, at the same functional rates as exact computation. We consider fixed- and high-dimensional problems, group and hierarchical constraints, and the possibility that all models are misspecified. We also obtain ALA rates for Gaussian regression under non-local priors, an important example where the LA can be costly and does not consistently estimate the integrated likelihood. Our examples include non-linear regression, logistic, Poisson and survival models. We implement the methodology in the R package mombf.

Online hierarchical forecasting for power consumption data

This paper proposes a three-step approach to forecasting time series of electricity consumption at different levels of household aggregation. These series are linked by hierarchical constraints—global consumption is the sum of regional consumption, for example. First, benchmark forecasts are generated for all series using generalized additive models. Second, for each series, the aggregation algorithm ML-Poly, introduced by Gaillard, Stoltz, and van Erven in 2014, finds an optimal linear combination of the benchmarks. Finally, the forecasts are projected onto a coherent subspace to ensure that the final forecasts satisfy the hierarchical constraints. By minimizing a regret criterion, we show that the aggregation and projection steps improve the root mean square error of the forecasts. Our approach is tested on household electricity consumption data; experimental results suggest that successive aggregation and projection steps improve the benchmark forecasts at different levels of household aggregation.

Multi-Directional Multi-Label Learning

In multi-label learning, the key problem is to capture the relationships between multiple labels, including proximities and unconformities. In this paper, we consider the relationships among multiple labels from multi-directions, including utilizing discriminative classifier, proposing a general hierarchical constraint and proximity correlation, meanwhile combining low-rank constraint, to infer a novel Multi-Directional Multi-Label learning (MDML) model. To optimize the problems involved in to the proposed models, we develop an iterative algorithms based on the alternating direction method of multipliers (ADMM) algorithm. In the simulations, the experimental results on 4 popular benchmark datasets demonstrate the superiorities of MDML model.

A hierarchical constraint approach for dynamic modeling and trajectory tracking control of a mobile robot

The dynamic modeling and trajectory tracking control of a mobile robot is handled by a hierarchical constraint approach in this study. When the wheeled mobile robot with complex generalized coordinates has structural constraints and motion constraints, the number of constraints is large and the properties of them are different. Therefore, it is difficult to get the dynamic model and trajectory tracking control force of the wheeled mobile robot at the same time. To solve the aforementioned problem, a creative hierarchical constraint approach based on the Udwadia–Kalaba theory is proposed. In this approach, constraints are classified into two levels, structural constraints are the first level and motion constraints are the second level. In the second level constraint, arbitrary initial conditions may cause the trajectory to diverge. Thus, we propose the asymptotic convergence criterion to deal with it. Then, the analytical dynamic equation and trajectory tracking control force of the wheeled mobile robot can be obtained simultaneously. To verify the effectiveness and accuracy of this methodology, a numerical simulation of a three-wheeled mobile robot is carried out.

Perceived barriers faced by Chinese adult children in social support of their elderly parents' outbound travels

This study aimed to investigate the perceived barriers faced by Chinese adult children in supporting their parents' outbound travel. Semi-structured in-depth interviews using thematic analysis were conducted with 21 Chinese adult children. From the perspective of support providers (i.e., Chinese adult children), seven barriers of the perceived support barriers were identified: safety and health concerns, fear of disappointing families, unwillingness to scarify their leisure time and travel needs, perceived negative parental attitudes towards travelling abroad, high cost of raising a child, complicated border exit procedures, and lack of time. Filial piety as an important cultural background plays the role of an ‘enabler’ in realising the social support of adult children for their parents' outbound travel. The findings test the leisure constraint theory in Chinese cultural backgrounds and verify the applicability of the hierarchical constraints model to adult children support providers influenced by Confucian cultural values. Both theoretical and practical implications are discussed.