Hierarchical Prediction Research Articles

Hierarchical Prediction Based on Network-Representation-Learning-Enhanced Clustering for Bike-Sharing System in Smart City

The sharing bike system is emerging as a new type of transportation with the advance of the smart city in recent years. More and more people will choose to ride a sharing bicycle for short-distance travel. While sharing bikes provide convenient services to customers, there are also many unfavorable factors in the sharing bike system, which have a certain impact on the riding experience of customers. One of the obvious disadvantages is the imbalance of usage because of the abnormal distribution of sharing bike stations in different areas. Despite some prediction work have been done, most of them merely consider the geographical factors between station whereas the user preferences and global network information are not fully considered. In this work, we propose a hierarchical model for sharing bike prediction, which can predict the number of rents/returns of each sharing bike station in the future to achieve resource redistribution. The proposed model is composed of two steps, including the sharing bike station clustering via network representation learning by considering the migration trends and geographic location information of bike-sharing between station. In the hierarchical prediction step, the total number of all bike-sharing station is predicted with an inference model based on multiple similarities. Finally, the number of rents/returns at each station can be derived. Our method is evaluated based on two publicly available sharing data sets by comparison with several baseline methods. Extensive experimental results on two open sharing bike data sets demonstrate that our proposed hierarchical can achieve the best prediction results. Our proposed method has a 17.34% improvement in root-mean-squared logarithmic error and a 10.4% improvement in return prediction.

Prediction of Knowledge Management for Success of Franchise Hospitality in a Post-Pandemic Economy

Due to its unpredictability, the novel coronavirus (COVID-19) pandemic has changed the global business climate and commercial management practices in unprecedented ways. As a direct result of the pandemic, the hospitality and tourism sectors have shut down, and business failure rates have occurred exponentially. The franchise hospitality industry has experienced significant impact and challenged a basic understanding of knowledge management (KM) implementation in the face of the COVID-19 outbreak. A strategic KM implementation practice can not only guide a large-scale operation, but also adjust an organization’s performance and competitiveness. The purpose of this study is to examine the influential criteria of success through effective KM implementation and to predict the probability of successful KM in a post-pandemic era. The conceptual framework for KM applies an analytic hierarchical prediction model reliant upon consistent fuzzy preference relations to assist the franchise hospitality sector’s consciousness of the influential criteria. An empirical case study is used to apply pairwise comparisons used to determine the priority weights and two possible outcomes. The case study will assist franchise organizations to analyze whether or not to implement KM, interdict application, or adopt revised actions. This assistance will enhance the success possibility of KM implementation within such a crisis environment. This study uses a case setting by assessing 15 franchises hospitality experts’ opinions in Taiwan relevant to KM implementation.

Formability prediction using bifurcation criteria and GTN damage model

In this paper, four plastic instability criteria, which are based on the bifurcation theory, are coupled with the GTN damage model for the prediction of diffuse and localized necking. General bifurcation (GB) criterion and limit-point bifurcation (LPB) criterion are used to predict diffuse necking, while loss of ellipticity (LOE) criterion and loss of strong ellipticity (LOSE) criterion are used to predict localized necking. The resulting constitutive equations and instability criteria are implemented into the finite element code ABAQUS/Standard. The constitutive equations are formulated within the framework of large deformations and fully three-dimensional approach. Since the developed numerical tools have intended applications mainly for thin sheet metals; therefore, the plane-stress conditions are considered within the instability criteria. The present contribution focuses on the effect of destabilizing mechanisms, due to non-associative plasticity and non-normal plastic flow rule, on the prediction of forming limit diagrams (FLDs). Theoretical classification of the bifurcation criteria, in terms of their order of prediction of critical necking strains, is first presented. Then, several variants of the GTN model are combined with the bifurcation criteria for the prediction of FLDs for fictitious materials. It is shown that the hierarchical prediction order of the different instability criteria is consistent with the theoretical classification, for all the considered variants of the GTN model. More specifically, it is shown that the GB criterion provides a lower bound to all bifurcation criteria, in terms of necking prediction, while the LOE criterion represents an upper bound.

Central Station-Based Demand Prediction for Determining Target Inventory in a Bike-Sharing System

Abstract Predicting the bike demand can help rebalance the bikes and improve the service quality of a bike-sharing system. A lot of works focus on predicting the bike demand for all the stations, which is unnecessary as the travel cost of rebalance operations increases sharply as the number of stations increases. In this paper, we propose a framework for predicting the hourly bike demand based on the central stations we define. Firstly, we propose Two-Stage Station Clustering Algorithm to assign central stations and common stations into each cluster. Secondly, we propose a hierarchical prediction model to predict the hourly bike demand for every cluster and each central station progressively. Thirdly, we use a well-studied queuing model to determine the target initial inventory for each central station. The most innovative contribution of this paper is proposing the concept of central station, the use of a novel algorithm to cluster the central stations and present a hierarchical model, containing the Time and Weather Similarity Weighted K-Nearest Neighbor Algorithm and a linear model to predict the bike demand for central stations. The experimental results on the New York citi bike system demonstrate that our proposed method is more accurate than other methods in solving existing problems.

Predictive processing in mental illness: Hierarchical circuitry for perception and trauma.

Predictive coding emerged as an explanation for how the brain can efficiently encode sensory stimuli. The hierarchical organization of neural circuits for perception thus passes prediction errors between computational layers. Extensions of this theory have provided a unifying understanding of Bayesian inference within neural circuits and psychiatric disorders. In particular, disorders of perception and belief have been explained as a Bayesian process of weighing prior beliefs (predictions) against new sensory data (prediction errors). The present issue of the Journal of Abnormal Psychology provides further evidence for how psychotic disorders develop and persist and how addiction- and trauma-related disorders may also be conceptualized. Trauma-related disorders in particular have begun to be identified as disorders of excessive accumulated prediction errors (uncertainty) over life. Here we summarize and reconcile recent advances in reinforcement learning momentum models with predictive processing and attempt to point out potential pitfalls for the application of hierarchical prediction models to stress. Future directions for understanding stress through this lens may need to involve updates to a purely hierarchical view or reframing long times cale molecular predictions as higher-order predictions. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Global Intake of Major Beverages in Adults by Country Wealth and Sociodemographic Characteristics: Analysis of the Global Dietary Database 2015

Abstract Objectives Assessing the impact of beverage intake on health has been limited by lack of reliable and standardized individual-level data on a global scale. We provide updated estimates of global intakes of sugar-sweetened beverages (SSBs), fruit juices, and milk, as well as novel estimates of coffee and tea, jointly by country, age, sex, education, and urban-rural residence for adults in 2015. Methods Beverge intake data in the Global Dietary Database were derived from 1144 national and subnational surveys from 1980 through 2015, covering 97.5% of the world's population. Mean intakes and 95% uncertainty intervals were estimated using a Bayesian hierarchical prediction model, and stratified by country (n = 185), urban-rural residence, sex, age group (all ages, 20 age groups), and education level (low, middle, high). The model combined primary data on individual-level intakes with time-varying country-level and survey-level covariates. Results Grouped by World Bank wealth categories, intake of SSBs was highest in low-income countries (LIC) and high-income countries (HIC) (0.7 servings (8-oz)/d each), and lowest in upper-middle income countries UMIC (0.4 s/d). Intakes were generally higher in men, urban residence, and with higher education, except in HIC were intakes did not differ by education. Fruit juice was highest in HIC (0.2 s/d) and lowest in lower-middle income countries (LMIC) (<0.04 s/d); and within countries, generally higher with urban residence and higher education. Coffee intake was highest in HIC (1.2 s/d) and lowest in LIC (0.7 s/d); and generally higher in men and with medium or high education level, without urban-rural differences. Tea intake was highest in HIC (0.8 s/d) and lowest in LMIC (0.5 s/d); and generally higher in women, urban residence, and with high education. Average milk intake ranged from 0.5 s/d in HIC to 0.2 s/d in LIC, LMIC and UMIC; and was generally higher at high education level, without differences by sex or urban-rural residence. Country-specific intakes and temporal trends since 1990 will be presented. Conclusions These novel global results highlight key findings and heterogeneity in intakes of different beverages, informing potential effects on health and corresponding policy priorities. Funding Sources Bill & Melinda Gates Foundation.

Global Plant-Based Food Intakes by Country Wealth and Socioeconomic Status: Findings from the Global Dietary Database

Abstract Objectives Given the major interest in plant-based foods (PF) for global diets, we aimed to characterize inequalities in intakes of fruits, vegetables (non-starchy, potatoes, other starchy), legumes, grains (refined, whole), and nuts/seeds by country wealth and population socioeconomic status (SES). Methods PF intakes were derived from the Global Dietary Database, including 1144 national and subnational surveys from 1980 through 2015 covering 97.5% of the world's population. A Bayesian hierarchical prediction model combined stratum-specific individual-level intakes with survey-level and time-varying country-level (GDP, FAO's food balance sheets) covariates to estimate mean intakes and 95% uncertainty intervals jointly stratified by country (n = 185), year (1990–2015), sex, age (all ages, 20 age groups), urban-rural residence, and education (low, middle, high). Results Using the World Bank's country wealth categories, fruit intake, in 2015, was much higher in high-income (HIC) (120 g/d) and upper-middle income countries (UMIC) (107 g/d) compared with low (LIC) (72 g/d) and lower-middle income countries (LMIC) (68 g/d); and within nations, generally higher with higher education and in urban areas, except for HIC where it was higher for rural residents. Average intakes of non-starchy vegetables (148–153 g/d) and legumes (22–25 g/d) were similar by country wealth; highest vegetable consumption was among those of higher education in LIC. Legume consumption was highest among those of higher education in LIC and LMIC. Average refined and whole grain intakes were highest in HIC (128 and 42 g/d) and lowest in LIC (46 and 11 g/d), with notable heterogeneity by education; refined grains were highest with lower education in HIC and UMIC, whereas whole grains were highest with higher education. Findings on other PF, and over time will be presented. Conclusions Such global data provide novel evidence for substantial variation in PF intakes by country wealth, further linked to population SES, and can inform potential health impacts and country-specific nutrition policy priorities. Funding Sources Gates Foundation.

Hierarchical Prediction and Adversarial Learning For Conditional Response Generation

There are a variety of underlying factors influencing what and how people communicate in their daily life. The ability to capture and utilize these factors enables the conversational systems to generate favorable responses and set up amicable connections with users. In this work, we investigate two major factors in response generation, i.e., emotion and intention. To explore the dependency between them, we develop a hierarchical variational model that predicts in sequence the emotion and intention to be conveyed in a response. The response can then be generated word-by-word based on the predictions. We also apply a novel adversarial-augmented inference network to facilitate model training. The experimental results demonstrate the effectiveness of the proposed model as well as the novel adversarial objective. The hypothesis that emotion shapes human communication behavior is also validated.

Identification and Validation of an Immune-Related Gene-Based Prognostic Index for Bladder Cancer

Background: Bladder cancer (BLCA) is a common malignant tumour of the urinary system. Knowledge of the mechanism of and treatment methods for BLCA remain limited. Methods: Robust rank aggregation was used to integrate five GEO bladder cancer microarray datasets for identifying differentially expressed genes (DEGs) between NMIBC and MIBC. An immune-related gene (IRG)-based prognostic index (IRGPI) was constructed using multivariate analysis. The reliability of the IRGPI was evaluated using independent datasets, and correlations between the IRGPI and clinicopathological characteristics as well as the immune microenvironment were evaluated. Finally, a nomogram was established to evaluate patient prognosis with BLCA. Findings: We identified 154 DEGs related to the degree of BLCA infiltration, including 24 IRGs. Missense mutation was the most common mutation type in IRGs. Ten hub IRGs were identified by protein-protein interaction network analysis. GSEA and GSVA of two novel BLCA-related genes (TYROBP and FCER1G) revealed that they were related to immunity. Nine survival-related IRGs were identified and their potential regulation by transcription factors was analysed. An independent IRGPI (CTSE, CXCL10, FAM3B, MMP9, OLR1, and S100P) was established, and was confirmed to be related to tumour stage, tumour grade, T stage, and N stage. The nomogram could effectively predict the prognosis of BLCA patients. Importantly, the IRGPI reflected infiltration by several types of immune cells. Interpretation: Our data provide new insights in the pathogenesis of BLCA and target genes for immunotherapy. The application of molecular markers for hierarchical prediction paves the way to precision medicine. Funding: This study was financially supported by the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (grant number: 2018-I2M-1-002) and the Beijing Hospital Clinical Research 121 Project (BJ-2018-090). Declaration of Interest: The authors have declared no competing interests.

Hierarchical Intrusion Detection Using Machine Learning and Knowledge Model

Intrusion detection systems (IDS) present a critical component of network infrastructures. Machine learning models are widely used in the IDS to learn the patterns in the network data and to detect the possible attacks in the network traffic. Ensemble models combining a variety of different machine learning models proved to be efficient in this domain. On the other hand, knowledge models have been explicitly designed for the description of the attacks and used in ontology-based IDS. In this paper, we propose a hierarchical IDS based on the original symmetrical combination of machine learning approach with knowledge-based approach to support detection of existing types and severity of new types of network attacks. Multi-stage hierarchical prediction consists of the predictive models able to distinguish the normal connections from the attacks and then to predict the attack classes and concrete attack types. The knowledge model enables to navigate through the attack taxonomy and to select the appropriate model to perform a prediction on the selected level. Designed IDS was evaluated on a widely used KDD 99 dataset and compared to similar approaches.

A Tale of Two Positivities and the N400: Distinct Neural Signatures Are Evoked by Confirmed and Violated Predictions at Different Levels of Representation.

It has been proposed that hierarchical prediction is a fundamental computational principle underlying neurocognitive processing. Here, we ask whether the brain engages distinct neurocognitive mechanisms in response to inputs that fulfill versus violate strong predictions at different levels of representation during language comprehension. Participants read three-sentence scenarios in which the third sentence constrained for a broad event structure, for example, {Agent caution animate-Patient}. High constraint contexts additionally constrained for a specific event/lexical item, for example, a two-sentence context about a beach, lifeguards, and sharks constrained for the event, {Lifeguards cautioned Swimmers}, and the specific lexical item swimmers. Low constraint contexts did not constrain for any specific event/lexical item. We measured ERPs on critical nouns that fulfilled and/or violated each of these constraints. We found clear, dissociable effects to fulfilled semantic predictions (a reduced N400), to event/lexical prediction violations (an increased late frontal positivity), and to event structure/animacy prediction violations (an increased late posterior positivity/P600). We argue that the late frontal positivity reflects a large change in activity associated with successfully updating the comprehender's current situation model with new unpredicted information. We suggest that the late posterior positivity/P600 is triggered when the comprehender detects a conflict between the input and her model of the communicator and communicative environment. This leads to an initial failure to incorporate the unpredicted input into the situation model, which may be followed by second-pass attempts to make sense of the discourse through reanalysis, repair, or reinterpretation. Together, these findings provide strong evidence that confirmed and violated predictions at different levels of representation manifest as distinct spatiotemporal neural signatures.

Adaptive Lagrangian Multiplier for Quantization Parameter Cascading in HEVC Hierarchical Coding

Hierarchical prediction structures applied in High Efficiency Video Coding (H.265/HEVC) leads to significant coding efficiency improvement. Quantization parameter cascading (QPC) techniques designed for specific prediction structure employs different quantization intervals for different temporal layers that further improve rate-distortion (R-D) performance of hierarchical coding. The existing QPC schemes focus on quantization parameter allocation, however, ignore the adaptation of Lagrangian multiplier to prediction structure, inter-layer residual dependency and layer-different quantization parameters (QPs). The misalignment of employed Lagrangian multiplier to the optimal one restricts the performance of QPC methods. To promote QPC schemes to their best performance, we propose a Lagrangian multiplier adaptation method based on accurate hyperbolic R-D model, distortion-QP model and rate-QP model. The derived Lagrangian multiplier is adaptive to residual characteristics under different prediction structures and QPC methods. Experimental results show that our proposed Lagrangian multipliers applied to existing QPC methods achieves Bjontegaard delta bit rate (BDBR) gain of 3.59 ${\%}$ to 4.65 ${\%}$ with small computational complexity increase.

Rate Distortion Optimization: A Joint Framework and Algorithms for Random Access Hierarchical Video Coding.

This paper revisits the problem of rate distortion optimization (RDO) with focus on inter-picture dependence. A joint RDO framework which incorporates the Lagrange multiplier as one of parameters to be optimized is proposed. Simplification strategies are demonstrated for practical applications. To make the problem tractable, we consider an approach where prediction residuals of pictures in a video sequence are assumed to be emitted from a finite set of sources. Consequently the RDO problem is formulated as finding optimal coding parameters for a finite number of sources, regardless of the length of the video sequence. Specifically, in cases where a hierarchical prediction structure is used, prediction residuals of pictures at the same prediction layer are assumed to be emitted from a common source. Following this approach, we propose an iterative algorithm to alternatively optimize the selections of quantization parameters (QPs) and the corresponding Lagrange multipliers. Based on the results of the iterative algorithm, we further propose two practical algorithms to compute QPs and the Lagrange multipliers for the RA(random access) hierarchical video coding: the first practical algorithm uses a fixed formula to compute QPs and the Lagrange multipliers, and the second practical algorithm adaptively adjusts both QPs and the Lagrange multipliers. Experimental results show that these three algorithms, integrated into the HM 16.20 reference software of HEVC, can achieve considerable RD improvements over the standard HM 16.20 encoder, in the common RA test configuration.

Sub-Frame Layer Rate Distortion Optimization for High Efficiency Video Coding

This paper proposes a sub-frame layer RDO algorithm to further improve coding performance of low-delay hierarchical prediction structure in High Efficiency Video Coding. In the proposed algorithm, each frame is segmented into two sub-frames according to motion intensity, and a sub-frame layer is taken as a basic optimizing unit. An inter-frame dependency model is proposed to quantify inter-frame dependency of a frame or a sub-frame. The proposed inter-frame dependency model is composed of two parts, Laplace distribution based direct inter-frame dependency model and distortion propagation chain. The direct inter-frame dependency model is used to evaluate dependency caused by direct reference, and the distortion propagation chain shows the relationship between direct reference and indirect reference. With quantified inter-frame dependency, Lagrange optimization is then used to solve the optimal $\lambda $ -QP pair for each sub-frame layer in frame layer 1 (L1) to frame layer 4 (L4). For frame layer 0 (L0), additional L0 frames are adaptively inserted according to inter-frame dependency of the previous L0 frame, and the optimal $\lambda $ -QP pair for each sub-frame in L0 is solved using Lagrange optimization. When peak signal to noise ratio is used as quality criterion, the proposed algorithm achieves 7.0% and 7.0% bit-rate reduction with lowdelay_P_main (LDP) and lowdelay_main (LDB) configurations respectively; when structural similarity is used as quality criterion, the proposed algorithm achieves 13.3% and 13.1% bit-rate reduction with LDP and LDB configurations respectively.

Minimizing prediction errors in predictive processing: from inconsistency to non-representationalism

Predictive processing is an increasingly popular approach to cognition, perception and action. It says that the brain is essentially a hierarchical prediction machine. It is typically construed in a representationalist and inferentialist fashion so that the brain makes contentful inferences on the basis of representational models. In this paper, I argue that the predictive processing framework is inconsistent with this epistemic position. In particular, I argue that the combination of hierarchical modeling, contentful inferentialism and representationalism entail an internal inconsistency. Specifically, for a particular set of states, there will be both a representation requirement and not. Yet a system cannot both be required to represent a certain set of states and not be required to represent those states. Due to this contradiction, I propose to reject the standard view. I suggest that predictive processing is best interpreted in terms of reliable covariation instead, entailing an instrumentalist approach to the statistical machinery.

Features-Enhanced Multi-Attribute Estimation with Convolutional Tensor Correlation Fusion Network

To achieve robust facial attribute estimation, a hierarchical prediction system referred to as tensor correlation fusion network (TCFN) is proposed for attribute estimation. The system includes feature extraction, correlation excavation among facial attribute features, score fusion, and multi-attribute prediction. Subnetworks (Age-Net, Gender-Net, Race-Net, and Smile-Net) are used to extract corresponding features while Main-Net extracts features not only from an input image but also from corresponding pooling layers of subnetworks. Dynamic tensor canonical correlation analysis (DTCCA) is proposed to explore the correlation of different targets’ features in the F7 layers. Then, for binary classifications of gender, race, and smile, corresponding robust decisions are achieved by fusing the results of subnetworks with those of TCFN while for age prediction, facial image into one of age groups, and then ELM regressor performs the final age estimation. Experimental results on benchmarks with multiple face attributes (MORPH-II, Adience Benchmark datasets, LAP-2016, and CelebA) show that the proposed approach has superior performance compared to state of the art.

Technique for lossless compression of color images based on hierarchical prediction, inversion, and context adaptive coding

Among the variety of multimedia formats, color images play a prominent role. A technique for lossless compression of color images is introduced. The technique is composed of first transforming a red, green, and blue image into luminance and chrominance domain (Y C u C v). Then, the luminance channel Y is compressed with a context-based, adaptive, lossless image coding technique (CALIC). After processing the chrominance channels with a hierarchical prediction technique that was introduced earlier, Burrows–Wheeler inversion coder or JPEG 2000 is used in compression of those C u and C v channels. It is demonstrated that, on a wide variety of images, particularly on medical images, the technique achieves substantial compression gains over other well-known compression schemes, such as CALIC, M-CALIC, Better Portable Graphics, JPEG-LS, JPEG 2000, and the previously proposed hierarchical prediction and context-adaptive coding technique LCIC.

Implementation of Block-Based Hierarchical Prediction for Developing an Error-Propagation-Free Reversible Data Hiding Scheme

This paper proposes a reversible data hiding technique based on the residual histogram shifting technique. To improve the hiding capacity, this study proposes a multiple-round hierarchical prediction mechanism that generates the prediction errors of each image block. The prediction errors of each block are collected to produce the residual histogram and the secret data are then embedded into the residual histogram to obtain the embedded image. Experimental results demonstrate that the proposed technique not only provides good hiding capacity, but also maintains good image quality of the embedded image. In addition, this technique can be easily extended for image integrity protection as it is capable of resisting error propagation.

Predict Consumers’ Automobile Purchase Behavior Based on the Preferences of Explicit Features and Implicit Topic

In order to know the role of online reviews in predicting consumers’ automobile purchasing behavior in the social network environment, this paper uses online reviews to construct a consumer interest preference model to obtain consumers’ explicit feature preferences and implicit topic preferences. Based on the preferences of consumers, introduce the tree structure to construct hierarchical prediction model to predict consumers’ purchase behavior. Meanwhile, consider the effect of online interactive comment on consumers’ automobile purchases. Several results are obtained in this paper:(1) The proposed hierarchical model introducing more information could achieve multi-granularity purchase-decision predicting;(2) the predictive effect based on implicit topic preference is better than that based on explicit feature preference; (3) integrating online interactive comments have positive effect in analyzing and predicting consumers’ purchasing behavior.

Non-iterative structural topology optimization using deep learning

This paper presents a non-iterative topology optimizer for conductive heat transfer structures with the help of deep learning. An artificial neural network is trained to deal with the black-and-white pixel images and generate near-optimal structures. Our design is a two-stage hierarchical prediction–refinement pipeline consisting of two coupled neural networks: a generative adversarial network (GAN) for predicting a low resolution near-optimal structure and a super-resolution generative adversarial network (SRGAN) for predicting the refined structure in high resolution. Training datasets with given boundary conditions and the optimized pixel image structures are obtained after simulating a big amount of topology optimization procedures. For more effective training and inference, these datasets are generated with two different resolutions. Experiments demonstrated that our learning based optimizer can provide accurate estimation of the conductive heat transfer topology using negligible computational time. This effective incorporation of deep learning into topology optimization could enable promising applications in large-scale engineering structure design.