Academic literature on the topic 'Large dimensional learning'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the lists of relevant articles, books, theses, conference reports, and other scholarly sources on the topic 'Large dimensional learning.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Journal articles on the topic "Large dimensional learning"
1
Khan, Usman A., Soummya Kar, and José M. F. Moura. "Higher Dimensional Consensus: Learning in Large-Scale Networks." IEEE Transactions on Signal Processing 58, no. 5 (May 2010): 2836–49. http://dx.doi.org/10.1109/tsp.2010.2042482.
Full text
2
Lin, Zhiping, Jiuwen Cao, Tao Chen, Yi Jin, Zhan-Li Sun, and Amaury Lendasse. "Extreme Learning Machine on High Dimensional and Large Data Applications." Mathematical Problems in Engineering 2015 (2015): 1–2. http://dx.doi.org/10.1155/2015/624903.
Full text
3
Peng, Chong, Jie Cheng, and Qiang Cheng. "A Supervised Learning Model for High-Dimensional and Large-Scale Data." ACM Transactions on Intelligent Systems and Technology 8, no. 2 (January 18, 2017): 1–23. http://dx.doi.org/10.1145/2972957.
Full text
4
Terol, Rafael Munoz, Alejandro Reina Reina, Saber Ziaei, and David Gil. "A Machine Learning Approach to Reduce Dimensional Space in Large Datasets." IEEE Access 8 (2020): 148181–92. http://dx.doi.org/10.1109/access.2020.3012836.
Full text
5
Keriven, Nicolas, Anthony Bourrier, Rémi Gribonval, and Patrick Pérez. "Sketching for large-scale learning of mixture models." Information and Inference: A Journal of the IMA 7, no. 3 (December 22, 2017): 447–508. http://dx.doi.org/10.1093/imaiai/iax015.
Full textAbstract:
Abstract Learning parameters from voluminous data can be prohibitive in terms of memory and computational requirements. We propose a ‘compressive learning’ framework, where we estimate model parameters from a sketch of the training data. This sketch is a collection of generalized moments of the underlying probability distribution of the data. It can be computed in a single pass on the training set and is easily computable on streams or distributed datasets. The proposed framework shares similarities with compressive sensing, which aims at drastically reducing the dimension of high-dimensional signals while preserving the ability to reconstruct them. To perform the estimation task, we derive an iterative algorithm analogous to sparse reconstruction algorithms in the context of linear inverse problems. We exemplify our framework with the compressive estimation of a Gaussian mixture model (GMM), providing heuristics on the choice of the sketching procedure and theoretical guarantees of reconstruction. We experimentally show on synthetic data that the proposed algorithm yields results comparable to the classical expectation-maximization technique while requiring significantly less memory and fewer computations when the number of database elements is large. We further demonstrate the potential of the approach on real large-scale data (over $10^{8}$ training samples) for the task of model-based speaker verification. Finally, we draw some connections between the proposed framework and approximate Hilbert space embedding of probability distributions using random features. We show that the proposed sketching operator can be seen as an innovative method to design translation-invariant kernels adapted to the analysis of GMMs. We also use this theoretical framework to derive preliminary information preservation guarantees, in the spirit of infinite-dimensional compressive sensing.
6
Panos, Aristeidis, Petros Dellaportas, and Michalis K. Titsias. "Large scale multi-label learning using Gaussian processes." Machine Learning 110, no. 5 (April 14, 2021): 965–87. http://dx.doi.org/10.1007/s10994-021-05952-5.
Full textAbstract:
AbstractWe introduce a Gaussian process latent factor model for multi-label classification that can capture correlations among class labels by using a small set of latent Gaussian process functions. To address computational challenges, when the number of training instances is very large, we introduce several techniques based on variational sparse Gaussian process approximations and stochastic optimization. Specifically, we apply doubly stochastic variational inference that sub-samples data instances and classes which allows us to cope with Big Data. Furthermore, we show it is possible and beneficial to optimize over inducing points, using gradient-based methods, even in very high dimensional input spaces involving up to hundreds of thousands of dimensions. We demonstrate the usefulness of our approach on several real-world large-scale multi-label learning problems.
7
Cao, Jiuwen, and Zhiping Lin. "Extreme Learning Machines on High Dimensional and Large Data Applications: A Survey." Mathematical Problems in Engineering 2015 (2015): 1–13. http://dx.doi.org/10.1155/2015/103796.
Full textAbstract:
Extreme learning machine (ELM) has been developed for single hidden layer feedforward neural networks (SLFNs). In ELM algorithm, the connections between the input layer and the hidden neurons are randomly assigned and remain unchanged during the learning process. The output connections are then tuned via minimizing the cost function through a linear system. The computational burden of ELM has been significantly reduced as the only cost is solving a linear system. The low computational complexity attracted a great deal of attention from the research community, especially for high dimensional and large data applications. This paper provides an up-to-date survey on the recent developments of ELM and its applications in high dimensional and large data. Comprehensive reviews on image processing, video processing, medical signal processing, and other popular large data applications with ELM are presented in the paper.
8
Ju, Cheng, Susan Gruber, Samuel D. Lendle, Antoine Chambaz, Jessica M. Franklin, Richard Wyss, Sebastian Schneeweiss, and Mark J. van der Laan. "Scalable collaborative targeted learning for high-dimensional data." Statistical Methods in Medical Research 28, no. 2 (September 22, 2017): 532–54. http://dx.doi.org/10.1177/0962280217729845.
Full textAbstract:
Robust inference of a low-dimensional parameter in a large semi-parametric model relies on external estimators of infinite-dimensional features of the distribution of the data. Typically, only one of the latter is optimized for the sake of constructing a well-behaved estimator of the low-dimensional parameter of interest. Optimizing more than one of them for the sake of achieving a better bias-variance trade-off in the estimation of the parameter of interest is the core idea driving the general template of the collaborative targeted minimum loss-based estimation procedure. The original instantiation of the collaborative targeted minimum loss-based estimation template can be presented as a greedy forward stepwise collaborative targeted minimum loss-based estimation algorithm. It does not scale well when the number p of covariates increases drastically. This motivates the introduction of a novel instantiation of the collaborative targeted minimum loss-based estimation template where the covariates are pre-ordered. Its time complexity is [Formula: see text] as opposed to the original [Formula: see text], a remarkable gain. We propose two pre-ordering strategies and suggest a rule of thumb to develop other meaningful strategies. Because it is usually unclear a priori which pre-ordering strategy to choose, we also introduce another instantiation called SL-C-TMLE algorithm that enables the data-driven choice of the better pre-ordering strategy given the problem at hand. Its time complexity is [Formula: see text] as well. The computational burden and relative performance of these algorithms were compared in simulation studies involving fully synthetic data or partially synthetic data based on a real world large electronic health database; and in analyses of three real, large electronic health databases. In all analyses involving electronic health databases, the greedy collaborative targeted minimum loss-based estimation algorithm is unacceptably slow. Simulation studies seem to indicate that our scalable collaborative targeted minimum loss-based estimation and SL-C-TMLE algorithms work well. All C-TMLEs are publicly available in a Julia software package.
9
Loyola R, Diego G., Mattia Pedergnana, and Sebastián Gimeno García. "Smart sampling and incremental function learning for very large high dimensional data." Neural Networks 78 (June 2016): 75–87. http://dx.doi.org/10.1016/j.neunet.2015.09.001.
Full text
10
Tran, Loc, Debrup Banerjee, Jihong Wang, Ashok J. Kumar, Frederic McKenzie, Yaohang Li, and Jiang Li. "High-dimensional MRI data analysis using a large-scale manifold learning approach." Machine Vision and Applications 24, no. 5 (April 19, 2013): 995–1014. http://dx.doi.org/10.1007/s00138-013-0499-8.
Full textDissertations / Theses on the topic "Large dimensional learning"
1
Bussy, Simon. "Introduction of high-dimensional interpretable machine learning models and their applications." Thesis, Sorbonne université, 2019. http://www.theses.fr/2019SORUS488.
Full textAbstract:
Dans ce manuscrit sont introduites de nouvelles méthodes interprétables de machine learning dans un contexte de grande dimension. Différentes procédures sont alors proposées : d'abord le C-mix, un modèle de mélange de durées qui détecte automatiquement des sous-groupes suivant le risque d'apparition rapide de l'événement temporel étudié; puis la pénalité binarsity, une combinaison entre variation totale pondérée et contrainte linéaire par bloc qui s'applique sur l'encodage "one-hot'' de covariables continues ; et enfin la méthode binacox qui applique la pénalité précédente dans un modèle de Cox en tirant notamment parti de sa propriété de détection automatique de seuils dans les covariables continues. Pour chacune d'entre elles, les propriétés théoriques sont étudiées comme la convergence algorithmique ou l'établissement d'inégalités oracles non-asymptotiques, et une étude comparative avec l'état de l'art est menée sur des données simulées et réelles. Toutes les méthodes obtiennent de bons résultats prédictifs ainsi qu'en terme de complexité algorithmique, et chacune dispose d'atouts intéressants sur le plan de l'interprétabilitéThis dissertation focuses on the introduction of new interpretable machine learning methods in a high-dimensional setting. We developped first the C-mix, a mixture model of censored durations that automatically detects subgroups based on the risk that the event under study occurs early; then the binarsity penalty combining a weighted total variation penalty with a linear constraint per block, that applies on one-hot encoding of continuous features; and finally the binacox model that uses the binarsity penalty within a Cox model to automatically detect cut-points in the continuous features. For each method, theoretical properties are established: algorithm convergence, non-asymptotic oracle inequalities, and comparison studies with state-of-the-art methods are carried out on both simulated and real data. All proposed methods give good results in terms of prediction performances, computing time, as well as interpretability abilities
2
Rawald, Tobias. "Scalable and Efficient Analysis of Large High-Dimensional Data Sets in the Context of Recurrence Analysis." Doctoral thesis, Humboldt-Universität zu Berlin, 2018. http://dx.doi.org/10.18452/18797.
Full textAbstract:
Die Recurrence Quantification Analysis (RQA) ist eine Methode aus der nicht-linearen Zeitreihenanalyse. Im Mittelpunkt dieser Methode steht die Auswertung des Inhalts sogenannter Rekurrenzmatrizen. Bestehende Berechnungsansätze zur Durchführung der RQA können entweder nur Zeitreihen bis zu einer bestimmten Länge verarbeiten oder benötigen viel Zeit zur Analyse von sehr langen Zeitreihen. Diese Dissertation stellt die sogenannte skalierbare Rekurrenzanalyse (SRA) vor. Sie ist ein neuartiger Berechnungsansatz, der eine gegebene Rekurrenzmatrix in mehrere Submatrizen unterteilt. Jede Submatrix wird von einem Berechnungsgerät in massiv-paralleler Art und Weise untersucht. Dieser Ansatz wird unter Verwendung der OpenCL-Schnittstelle umgesetzt. Anhand mehrerer Experimente wird demonstriert, dass SRA massive Leistungssteigerungen im Vergleich zu existierenden Berechnungsansätzen insbesondere durch den Einsatz von Grafikkarten ermöglicht. Die Dissertation enthält eine ausführliche Evaluation, die den Einfluss der Anwendung mehrerer Datenbankkonzepte, wie z.B. die Repräsentation der Eingangsdaten, auf die RQA-Verarbeitungskette analysiert. Es wird untersucht, inwiefern unterschiedliche Ausprägungen dieser Konzepte Einfluss auf die Effizienz der Analyse auf verschiedenen Berechnungsgeräten haben. Abschließend wird ein automatischer Optimierungsansatz vorgestellt, der performante RQA-Implementierungen für ein gegebenes Analyseszenario in Kombination mit einer Hardware-Plattform dynamisch bestimmt. Neben anderen Aspekten werden drastische Effizienzgewinne durch den Einsatz des Optimierungsansatzes aufgezeigt.Recurrence quantification analysis (RQA) is a method from nonlinear time series analysis. It relies on the identification of line structures within so-called recurrence matrices and comprises a set of scalar measures. Existing computing approaches to RQA are either not capable of processing recurrence matrices exceeding a certain size or suffer from long runtimes considering time series that contain hundreds of thousands of data points. This thesis introduces scalable recurrence analysis (SRA), which is an alternative computing approach that subdivides a recurrence matrix into multiple sub matrices. Each sub matrix is processed individually in a massively parallel manner by a single compute device. This is implemented exemplarily using the OpenCL framework. It is shown that this approach delivers considerable performance improvements in comparison to state-of-the-art RQA software by exploiting the computing capabilities of many-core hardware architectures, in particular graphics cards. The usage of OpenCL allows to execute identical SRA implementations on a variety of hardware platforms having different architectural properties. An extensive evaluation analyses the impact of applying concepts from database technology, such memory storage layouts, to the RQA processing pipeline. It is investigated how different realisations of these concepts affect the performance of the computations on different types of compute devices. Finally, an approach based on automatic performance tuning is introduced that automatically selects well-performing RQA implementations for a given analytical scenario on specific computing hardware. Among others, it is demonstrated that the customised auto-tuning approach allows to considerably increase the efficiency of the processing by adapting the implementation selection.
3
Mai, Xiaoyi. "Méthodes des matrices aléatoires pour l’apprentissage en grandes dimensions." Thesis, Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLC078/document.
Full textAbstract:
Le défi du BigData entraîne un besoin pour les algorithmes d'apprentissage automatisé de s'adapter aux données de grande dimension et de devenir plus efficace. Récemment, une nouvelle direction de recherche est apparue qui consiste à analyser les méthodes d’apprentissage dans le régime moderne où le nombre n et la dimension p des données sont grands et du même ordre. Par rapport au régime conventionnel où n>>p, le régime avec n,p sont grands et comparables est particulièrement intéressant, car les performances d’apprentissage dans ce régime restent sensibles à l’ajustement des hyperparamètres, ouvrant ainsi une voie à la compréhension et à l’amélioration des techniques d’apprentissage pour ces données de grande dimension.L'approche technique de cette thèse s'appuie sur des outils avancés de statistiques de grande dimension, nous permettant de mener des analyses allant au-delà de l'état de l’art. La première partie de la thèse est consacrée à l'étude de l'apprentissage semi-supervisé sur des grandes données. Motivés par nos résultats théoriques, nous proposons une alternative supérieure à la méthode semi-supervisée de régularisation laplacienne. Les méthodes avec solutions implicites, comme les SVMs et la régression logistique, sont ensuite étudiées sous des modèles de mélanges réalistes, fournissant des détails exhaustifs sur le mécanisme d'apprentissage. Plusieurs conséquences importantes sont ainsi révélées, dont certaines sont même en contradiction avec la croyance communeThe BigData challenge induces a need for machine learning algorithms to evolve towards large dimensional and more efficient learning engines. Recently, a new direction of research has emerged that consists in analyzing learning methods in the modern regime where the number n and the dimension p of data samples are commensurately large. Compared to the conventional regime where n>>p, the regime with large and comparable n,p is particularly interesting as the learning performance in this regime remains sensitive to the tuning of hyperparameters, thus opening a path into the understanding and improvement of learning techniques for large dimensional datasets.The technical approach employed in this thesis draws on several advanced tools of high dimensional statistics, allowing us to conduct more elaborate analyses beyond the state of the art. The first part of this dissertation is devoted to the study of semi-supervised learning on high dimensional data. Motivated by our theoretical findings, we propose a superior alternative to the standard semi-supervised method of Laplacian regularization. The methods involving implicit optimizations, such as SVMs and logistic regression, are next investigated under realistic mixture models, providing exhaustive details on the learning mechanism. Several important consequences are thus revealed, some of which are even in contradiction with common belief
4
Chinot, Geoffrey. "Localization methods with applications to robust learning and interpolation." Electronic Thesis or Diss., Institut polytechnique de Paris, 2020. http://www.theses.fr/2020IPPAG002.
Full textAbstract:
Cette thèse de doctorat est centrée sur l'apprentissage supervisé. L'objectif principal est l'utilisation de méthodes de localisation pour obtenir des vitesses rapides de convergence, c'est-à-dire, des vitesse de l'ordre O(1/n), où n est le nombre d'observations. Ces vitesses ne sont pas toujours atteignables. Il faut imposer des contraintes sur la variance du problème comme une condition de Bernstein ou de marge. Plus particulièrement, dans cette thèse nous tentons d'établir des vitesses rapides de convergences pour des problèmes de robustesse et d'interpolation.On dit qu'un estimateur est robuste si ce dernier présente certaines garanties théoriques, sous le moins d'hypothèses possibles. Cette problématique de robustesse devient de plus en plus populaire. La raison principale est que dans l'ère actuelle du “big data", les données sont très souvent corrompues. Ainsi, construire des estimateurs fiables dans cette situation est essentiel. Dans cette thèse nous montrons que le fameux minimiseur du risque empirique (regularisé) associé à une fonction de perte Lipschitz est robuste à des bruits à queues lourde ainsi qu'a des outliers dans les labels. En revanche si la classe de prédicteurs est à queue lourde, cet estimateur n'est pas fiable. Dans ce cas, nous construisons des estimateurs appelé estimateur minmax-MOM, optimal lorsque les données sont à queues lourdes et possiblement corrompues.En apprentissage statistique, on dit qu'un estimateur interpole, lorsque ce dernier prédit parfaitement sur un jeu d'entrainement. En grande dimension, certains estimateurs interpolant les données peuvent être bons. En particulier, cette thèse nous étudions le modèle linéaire Gaussien en grande dimension et montrons que l'estimateur interpolant les données de plus petite norme est consistant et atteint même des vitesses rapidesThis PhD thesis deals with supervized machine learning and statistics. The main goal is to use localization techniques to derive fast rates of convergence, with a particular focus on robust learning and interpolation problems.Localization methods aim to analyze localized properties of an estimator to obtain fast rates of convergence, that is rates of order O(1/n), where n is the number of observations. Under assumptions, such as the Bernstein condition, such rates are attainable.A robust estimator is an estimator with good theoretical guarantees, under as few assumptions as possible. This question is getting more and more popular in the current era of big data. Large dataset are very likely to be corrupted and one would like to build reliable estimators in such a setting. We show that the well-known regularized empirical risk minimizer (RERM) with Lipschitz-loss function is robust with respect to heavy-tailed noise and outliers in the label. When the class of predictor is heavy-tailed, RERM is not reliable. In this setting, we show that minmax Median of Means estimators can be a solution. By construction minmax-MOM estimators are also robust to an adversarial contamination.Interpolation problems study learning procedure with zero training error. Surprisingly, in large dimension, interpolating the data does not necessarily implies over-fitting. We study a high dimensional Gaussian linear model and show that sometimes the over-fitting may be benign
5
Hmamouche, Youssef. "Prédiction des séries temporelles larges." Thesis, Aix-Marseille, 2018. http://www.theses.fr/2018AIXM0480.
Full textAbstract:
De nos jours, les systèmes modernes sont censés stocker et traiter des séries temporelles massives. Comme le nombre de variables observées augmente très rapidement, leur prédiction devient de plus en plus compliquée, et l’utilisation de toutes les variables pose des problèmes pour les modèles classiques.Les modèles de prédiction sans facteurs externes sont parmi les premiers modèles de prédiction. En vue d’améliorer la précision des prédictions, l’utilisation de multiples variables est devenue commune. Ainsi, les modèles qui tiennent en compte des facteurs externes, ou bien les modèles multivariés, apparaissent, et deviennent de plus en plus utilisés car ils prennent en compte plus d’informations.Avec l’augmentation des données liées entre eux, l’application des modèles multivariés devient aussi discutable. Le challenge dans cette situation est de trouver les facteurs les plus pertinents parmi l’ensemble des données disponibles par rapport à une variable cible.Dans cette thèse, nous étudions ce problème en présentant une analyse détaillée des approches proposées dans la littérature. Nous abordons le problème de réduction et de prédiction des données massives. Nous discutons également ces approches dans le contexte du Big Data.Ensuite, nous présentons une méthodologie complète pour la prédiction des séries temporelles larges. Nous étendons également cette méthodologie aux données très larges via le calcul distribué et le parallélisme avec une implémentation du processus de prédiction proposé dans l’environnement Hadoop/SparkNowadays, storage and data processing systems are supposed to store and process large time series. As the number of variables observed increases very rapidly, their prediction becomes more and more complicated, and the use of all the variables poses problems for classical prediction models.Univariate prediction models are among the first models of prediction. To improve these models, the use of multiple variables has become common. Thus, multivariate models and become more and more used because they consider more information.With the increase of data related to each other, the application of multivariate models is also questionable. Because the use of all existing information does not necessarily lead to the best predictions. Therefore, the challenge in this situation is to find the most relevant factors among all available data relative to a target variable.In this thesis, we study this problem by presenting a detailed analysis of the proposed approaches in the literature. We address the problem of prediction and size reduction of massive data. We also discuss these approaches in the context of Big Data.The proposed approaches show promising and very competitive results compared to well-known algorithms, and lead to an improvement in the accuracy of the predictions on the data used.Then, we present our contributions, and propose a complete methodology for the prediction of wide time series. We also extend this methodology to big data via distributed computing and parallelism with an implementation of the prediction process proposed in the Hadoop / Spark environment
6
Dang, Quang Vinh. "Évaluation de la confiance dans la collaboration à large échelle." Thesis, Université de Lorraine, 2018. http://www.theses.fr/2018LORR0002/document.
Full textAbstract:
Les systèmes collaboratifs à large échelle, où un grand nombre d’utilisateurs collaborent pour réaliser une tâche partagée, attirent beaucoup l’attention des milieux industriels et académiques. Bien que la confiance soit un facteur primordial pour le succès d’une telle collaboration, il est difficile pour les utilisateurs finaux d’évaluer manuellement le niveau de confiance envers chaque partenaire. Dans cette thèse, nous étudions le problème de l’évaluation de la confiance et cherchons à concevoir un modèle de confiance informatique dédiés aux systèmes collaboratifs. Nos travaux s’organisent autour des trois questions de recherche suivantes. 1. Quel est l’effet du déploiement d’un modèle de confiance et de la représentation aux utilisateurs des scores obtenus pour chaque partenaire ? Nous avons conçu et organisé une expérience utilisateur basée sur le jeu de confiance qui est un protocole d’échange d’argent en environnement contrôlé dans lequel nous avons introduit des notes de confiance pour les utilisateurs. L’analyse détaillée du comportement des utilisateurs montre que: (i) la présentation d’un score de confiance aux utilisateurs encourage la collaboration entre eux de manière significative, et ce, à un niveau similaire à celui de l’affichage du surnom des participants, et (ii) les utilisateurs se conforment au score de confiance dans leur prise de décision concernant l’échange monétaire. Les résultats suggèrent donc qu’un modèle de confiance peut être déployé dans les systèmes collaboratifs afin d’assister les utilisateurs. 2. Comment calculer le score de confiance entre des utilisateurs qui ont déjà collaboré ? Nous avons conçu un modèle de confiance pour les jeux de confiance répétés qui calcule les scores de confiance des utilisateurs en fonction de leur comportement passé. Nous avons validé notre modèle de confiance en relativement à: (i) des données simulées, (ii) de l’opinion humaine et (iii) des données expérimentales réelles. Nous avons appliqué notre modèle de confiance à Wikipédia en utilisant la qualité des articles de Wikipédia comme mesure de contribution. Nous avons proposé trois algorithmes d’apprentissage automatique pour évaluer la qualité des articles de Wikipédia: l’un est basé sur une forêt d’arbres décisionnels tandis que les deux autres sont basés sur des méthodes d’apprentissage profond. 3. Comment prédire la relation de confiance entre des utilisateurs qui n’ont pas encore interagi ? Etant donné un réseau dans lequel les liens représentent les relations de confiance/défiance entre utilisateurs, nous cherchons à prévoir les relations futures. Nous avons proposé un algorithme qui prend en compte les informations temporelles relatives à l’établissement des liens dans le réseau pour prédire la relation future de confiance/défiance des utilisateurs. L’algorithme proposé surpasse les approches de la littérature pour des jeux de données réels provenant de réseaux sociaux dirigés et signésLarge-scale collaborative systems wherein a large number of users collaborate to perform a shared task attract a lot of attention from both academic and industry. Trust is an important factor for the success of a large-scale collaboration. It is difficult for end-users to manually assess the trust level of each partner in this collaboration. We study the trust assessment problem and aim to design a computational trust model for collaborative systems. We focused on three research questions. 1. What is the effect of deploying a trust model and showing trust scores of partners to users? We designed and organized a user-experiment based on trust game, a well-known money-exchange lab-control protocol, wherein we introduced user trust scores. Our comprehensive analysis on user behavior proved that: (i) showing trust score to users encourages collaboration between them significantly at a similar level with showing nick- name, and (ii) users follow the trust score in decision-making. The results suggest that a trust model can be deployed in collaborative systems to assist users. 2. How to calculate trust score between users that experienced a collaboration? We designed a trust model for repeated trust game that computes user trust scores based on their past behavior. We validated our trust model against: (i) simulated data, (ii) human opinion, and (iii) real-world experimental data. We extended our trust model to Wikipedia based on user contributions to the quality of the edited Wikipedia articles. We proposed three machine learning approaches to assess the quality of Wikipedia articles: the first one based on random forest with manually-designed features while the other two ones based on deep learning methods. 3. How to predict trust relation between users that did not interact in the past? Given a network in which the links represent the trust/distrust relations between users, we aim to predict future relations. We proposed an algorithm that takes into account the established time information of the links in the network to predict future user trust/distrust relationships. Our algorithm outperforms state-of-the-art approaches on real-world signed directed social network datasets
7
Yang, Xiaoke. "Regularized Discriminant Analysis: A Large Dimensional Study." Thesis, 2018. http://hdl.handle.net/10754/627734.
Full textAbstract:
In this thesis, we focus on studying the performance of general regularized discriminant analysis (RDA) classifiers. The data used for analysis is assumed to follow Gaussian mixture model with different means and covariances. RDA offers a rich class of regularization options, covering as special cases the regularized linear discriminant analysis (RLDA) and the regularized quadratic discriminant analysis (RQDA) classi ers. We analyze RDA under the double asymptotic regime where the data dimension and the training size both increase in a proportional way. This double asymptotic regime allows for application of fundamental results from random matrix theory. Under the double asymptotic regime and some mild assumptions, we show that the asymptotic classification error converges to a deterministic quantity that only depends on the data statistical parameters and dimensions. This result not only implicates some mathematical relations between the misclassification error and the class statistics, but also can be leveraged to select the optimal parameters that minimize the classification error, thus yielding the optimal classifier. Validation results on the synthetic data show a good accuracy of our theoretical findings. We also construct a general consistent estimator to approximate the true classification error in consideration of the unknown previous statistics. We benchmark the performance of our proposed consistent estimator against classical estimator on synthetic data. The observations demonstrate that the general estimator outperforms others in terms of mean squared error (MSE).
8
Chajnacki, Gregory M. "Characteristics of learning organizations and multi-dimensional organizational performance indicators a survey of large, publicly-owned companies /." 2007. http://www.etda.libraries.psu.edu/theses/approved/WorldWideIndex/ETD-1775/index.html.
Full textBooks on the topic "Large dimensional learning"
1
Walker, Stephen G., and Mark Schafer. Operational Code Theory: Beliefs and Foreign Policy Decisions. Oxford University Press, 2018. http://dx.doi.org/10.1093/acrefore/9780190846626.013.411.
Full textAbstract:
The process of foreign policy decision making is influenced in large part by beliefs, along with the strategic interaction between actors engendered by their decisions and the resulting political outcomes. In this context, beliefs encompass three kinds of effects: the mirroring effects associated with the decision making situation, the steering effects that arise from this situation, and the learning effects of feedback. These effects are modeled using operational code analysis, although “operational code theory” more accurately describes an alliance of attribution and schema theories from psychology and game theory from economics applied to the domain of politics. This “theory complex” specifies belief-based solutions to the puzzles posed by diagnostic, decision making, and learning processes in world politics. The major social and intellectual dimensions of operational code theory can be traced to Nathan Leites’s seminal research on the Bolshevik operational code, The Operational Code of the Politburo. In the last half of the twentieth century, applications of operational code analysis have emphasized different cognitive, emotional, and motivational mechanisms as intellectual dimensions in explaining foreign policy decisions. The literature on operational code theory may be divided into four general waves of research: idiographic-interpretive studies, nomothetic-typological studies, quantitative-statistical studies, and formal modeling studies. The present trajectory of studies on operational code points to a number of important trends that straddle political psychology and game theory. For example, the psychological processes of mirroring, steering, and learning associated with operational code analysis have the potential to enrich our understanding of game-theoretic models of strategic interaction.
2
Sullivan, Mark, Nilay Patel, and Inderbir Gill. Principles of laparoscopic and robotic urological surgery. Edited by John Reynard. Oxford University Press, 2017. http://dx.doi.org/10.1093/med/9780199659579.003.0033.
Full textAbstract:
The development of laparoscopic and consequently robotic urological surgery have improved the visual field for the urological surgeon and led to reductions in postoperative pain, reduced convalescence, and improved cosmesis for the patient. Laparoscopy and robotics require video systems and telescopes to produce high-resolution images. Trocars have been developed to access the surgical field together with devices to deliver the insufflating gases. Instruments have been developed to allow for tissue dissection and incision together with haemostatic devices and sealants for control of small diameter vessel bleeding. Clips and staplers are used to control larger diameter vessels. Methods of access and skills training are discussed. Robotic surgery provides three-dimensional vision, greater range of movement, and the lack of tremor. Whether these are real benefits in terms of patient outcome is not yet clear, but the learning curve for robotic surgery does appear to be shorter than for pure laparoscopy.
3
Liang, Percy, Michael Jordan, and Dan Klein. Probabilistic grammars and hierarchical Dirichlet processes. Edited by Anthony O'Hagan and Mike West. Oxford University Press, 2018. http://dx.doi.org/10.1093/oxfordhb/9780198703174.013.27.
Full textAbstract:
This article focuses on the use of probabilistic context-free grammars (PCFGs) in natural language processing involving a large-scale natural language parsing task. It describes detailed, highly-structured Bayesian modelling in which model dimension and complexity responds naturally to observed data. The framework, termed hierarchical Dirichlet process probabilistic context-free grammar (HDP-PCFG), involves structured hierarchical Dirichlet process modelling and customized model fitting via variational methods to address the problem of syntactic parsing and the underlying problems of grammar induction and grammar refinement. The central object of study is the parse tree, which can be used to describe a substantial amount of the syntactic structure and relational semantics of natural language sentences. The article first provides an overview of the formal probabilistic specification of the HDP-PCFG, algorithms for posterior inference under the HDP-PCFG, and experiments on grammar learning run on the Wall Street Journal portion of the Penn Treebank.
4
Dobson, James E. Critical Digital Humanities. University of Illinois Press, 2019. http://dx.doi.org/10.5622/illinois/9780252042270.001.0001.
Full textAbstract:
This book seeks to develop an answer to the major question arising from the adoption of sophisticated data-science approaches within humanities research: are existing humanities methods compatible with computational thinking? Data-based and algorithmically powered methods present both new opportunities and new complications for humanists. This book takes as its founding assumption that the exploration and investigation of texts and data with sophisticated computational tools can serve the interpretative goals of humanists. At the same time, it assumes that these approaches cannot and will not obsolete other existing interpretive frameworks. Research involving computational methods, the book argues, should be subject to humanistic modes that deal with questions of power and infrastructure directed toward the field’s assumptions and practices. Arguing for a methodologically and ideologically self-aware critical digital humanities, the author contextualizes the digital humanities within the larger neo-liberalizing shifts of the contemporary university in order to resituate the field within a theoretically informed tradition of humanistic inquiry. Bringing the resources of critical theory to bear on computational methods enables humanists to construct an array of compelling and possible humanistic interpretations from multiple dimensions—from the ideological biases informing many commonly used algorithms to the complications of a historicist text mining, from examining the range of feature selection for sentiment analysis to the fantasies of human subjectless analysis activated by machine learning and artificial intelligence.
5
Austin, Kenneth. The Jews and the Reformation. Yale University Press, 2020. http://dx.doi.org/10.12987/yale/9780300186291.001.0001.
Full textAbstract:
This book examines the attitudes of various Christian groups in the Protestant and Catholic Reformations towards Jews, the Hebrew language, and Jewish learning. Martin Luther's writings are notorious, but Reformation attitudes were much more varied and nuanced than these might lead us to believe. The book has much to tell us about the Reformation and its priorities, and it has important implications for how we think about religious pluralism more broadly. The book begins by focusing on the impact and various forms of the Reformation on the Jews and pays close attention to the global perspective on Jewish experiences in the early modern period. It highlights the links between Jews in Europe and those in north Africa, Asia Minor, and the Americas, and it looks into the Jews' migrations and reputation as a corollary of Christians' exploration and colonisation of several territories. It seeks to next establish the position Jews occupied in Christian thinking and society by the start of the Reformation era, and then moves on to the first waves of reform in the earliest decades of the sixteenth century in both the Catholic and Protestant realms. The book explores the radical dimension to the Protestant Reformation and talks about identity as the heart of a fundamental issue associated with the Reformation. It analyzes “Counter Reformation” and discusses the various forms of Protestantism that had been accepted by large swathes of the population of many territories in Europe. Later chapters turn attention to relations between Jews and Christians in the first half of the seventeenth century and explore the Sabbatean movement as the most significant messianic movement since the first century BCE. In conclusion, the book summarizes how the Jews of Europe were in a very different position by the end of the seventeenth century compared to where they had been at the start of the sixteenth century. It recounts how Jewish communities sprung up in places which had not traditionally been a home to Jews, especially in Eastern Europe.
Book chapters on the topic "Large dimensional learning"
1
Lee, Sangkyun, and Andreas Holzinger. "Knowledge Discovery from Complex High Dimensional Data." In Solving Large Scale Learning Tasks. Challenges and Algorithms, 148–67. Cham: Springer International Publishing, 2016. http://dx.doi.org/10.1007/978-3-319-41706-6_7.
Full text
2
Rimal, Yagyanath. "Regression Analysis of Large Research Data: Dimensional Reduction Techniques." In Learning and Analytics in Intelligent Systems, 296–306. Cham: Springer International Publishing, 2020. http://dx.doi.org/10.1007/978-3-030-42363-6_35.
Full text
3
Shenoy, P. Deepa, K. G. Srinivasa, M. P. Mithun, K. R. Venugopal, and L. M. Patnaik. "Dynamic Subspace Clustering for Very Large High-Dimensional Databases." In Intelligent Data Engineering and Automated Learning, 850–54. Berlin, Heidelberg: Springer Berlin Heidelberg, 2003. http://dx.doi.org/10.1007/978-3-540-45080-1_117.
Full text
4
Wang, Dongxia, and Yongmei Lei. "Asynchronous Distributed ADMM for Learning with Large-Scale and High-Dimensional Sparse Data Set." In Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, 259–74. Cham: Springer International Publishing, 2019. http://dx.doi.org/10.1007/978-3-030-36405-2_27.
Full text
5
Zhang, Lijun, Tianbao Yang, Rong Jin, and Zhi-Hua Zhou. "Sparse Learning for Large-Scale and High-Dimensional Data: A Randomized Convex-Concave Optimization Approach." In Lecture Notes in Computer Science, 83–97. Cham: Springer International Publishing, 2016. http://dx.doi.org/10.1007/978-3-319-46379-7_6.
Full text
6
Qiu, Waishan, Wenjing Li, Xun Liu, and Xiaokai Huang. "Subjectively Measured Streetscape Qualities for Shanghai with Large-Scale Application of Computer Vision and Machine Learning." In Proceedings of the 2021 DigitalFUTURES, 242–51. Singapore: Springer Singapore, 2021. http://dx.doi.org/10.1007/978-981-16-5983-6_23.
Full textAbstract:
AbstractRecently, many new studies emerged to apply computer vision (CV) to street view imagery (SVI) dataset to objectively extract the view indices of various streetscape features such as trees to proxy urban scene qualities. However, human perceptions (e.g., imageability) have a subtle relationship to visual elements which cannot be fully captured using view indices. Conversely, subjective measures using survey and interview data explain more human behaviors. However, the effectiveness of integrating subjective measures with SVI dataset has been less discussed. To address this, we integrated crowdsourcing, CV, and machine learning (ML) to subjectively measure four important perceptions suggested by classical urban design theory. We first collected experts’ rating on sample SVIs regarding the four qualities which became the training labels. CV segmentation was applied to SVI samples extracting streetscape view indices as the explanatory variables. We then trained ML models and achieved high accuracy in predicting the scores. We found a strong correlation between predicted complexity score and the density of urban amenities and services Point of Interests (POI), which validates the effectiveness of subjective measures. In addition, to test the generalizability of the proposed framework as well as to inform urban renewal strategies, we compared the measured qualities in Pudong to other five renowned urban cores worldwide. Rather than predicting perceptual scores directly from generic image features using convolution neural network, our approach follows what urban design theory suggested and confirms various streetscape features affecting multi-dimensional human perceptions. Therefore, its result provides more interpretable and actionable implications for policymakers and city planners.
7
Goncalves, André R., Arindam Banerjee, Vidyashankar Sivakumar, and Soumyadeep Chatterjee. "Structured Estimation in High Dimensions." In Large-Scale Machine Learning in the Earth Sciences, 13–32. Boca Raton : Taylor & Francis, 2017. | Series: Chapman & Hall/CRC data mining & knowledge discovery series ; 42 | “A CRC title, part of the Taylor & Francis imprint, a member of the Taylor & Francis Group, the academic division of T&F Informa plc.”: Chapman and Hall/CRC, 2017. http://dx.doi.org/10.4324/9781315371740-2.
Full text
8
Behuet, Sabrina, Sebastian Bludau, Olga Kedo, Christian Schiffer, Timo Dickscheid, Andrea Brandstetter, Philippe Massicotte, Mona Omidyeganeh, Alan Evans, and Katrin Amunts. "A High-Resolution Model of the Human Entorhinal Cortex in the ‘BigBrain’ – Use Case for Machine Learning and 3D Analyses." In Lecture Notes in Computer Science, 3–21. Cham: Springer International Publishing, 2021. http://dx.doi.org/10.1007/978-3-030-82427-3_1.
Full textAbstract:
AbstractThe ‘BigBrain’ is a high-resolution data set of the human brain that enables three-dimensional (3D) analyses with a 20 µm spatial resolution at nearly cellular level. We use this data set to explore pre-α (cell) islands of layer 2 in the entorhinal cortex (EC), which are early affected in Alzheimer’s disease and have therefore been the focus of research for many years. They appear mostly in a round and elongated shape as shown in microscopic studies. Some studies suggested that islands may be interconnected based on analyses of their shape and size in two-dimensional (2D) space. Here, we characterized morphological features (shape, size, and distribution) of pre-α islands in the ‘BigBrain’, based on 3D-reconstructions of gapless series of cell-body-stained sections. The EC was annotated manually, and a machine-learning tool was trained to identify and segment islands with subsequent visualization using high-performance computing (HPC). Islands were visualized as 3D surfaces and their geometry was analyzed. Their morphology was complex: they appeared to be composed of interconnected islands of different types found in 2D histological sections of EC, with various shapes in 3D. Differences in the rostral-to-caudal part of EC were identified by specific distribution and size of islands, with implications for connectivity and function of the EC. 3D compactness analysis found more round and complex islands than elongated ones. The present study represents a use case for studying large microscopic data sets. It provides reference data for studies, e.g. investigating neurodegenerative diseases, where specific alterations in layer 2 were previously reported.
9
Habyarimana, Ephrem, and Sofia Michailidou. "Genomics Data." In Big Data in Bioeconomy, 69–76. Cham: Springer International Publishing, 2021. http://dx.doi.org/10.1007/978-3-030-71069-9_6.
Full textAbstract:
AbstractIn silico prediction of plant performance is gaining increasing breeders’ attention. Several statistical, mathematical and machine learning methodologies for analysis of phenotypic, omics and environmental data typically use individual or a few data layers. Genomic selection is one of the applications, where heterogeneous data, such as those from omics technologies, are handled, accommodating several genetic models of inheritance. There are many new high throughput Next Generation Sequencing (NGS) platforms on the market producing whole-genome data at a low cost. Hence, large-scale genomic data can be produced and analyzed enabling intercrosses and fast-paced recurrent selection. The offspring properties can be predicted instead of manually evaluated in the field . Breeders have a short time window to make decisions by the time they receive data, which is one of the major challenges in commercial breeding. To implement genomic selection routinely as part of breeding programs, data management systems and analytics capacity have therefore to be in order. The traditional relational database management systems (RDBMS), which are designed to store, manage and analyze large-scale data, offer appealing characteristics, particularly when they are upgraded with capabilities for working with binary large objects. In addition, NoSQL systems were considered effective tools for managing high-dimensional genomic data. MongoDB system, a document-based NoSQL database, was effectively used to develop web-based tools for visualizing and exploring genotypic information. The Hierarchical Data Format (HDF5), a member of the high-performance distributed file systems family, demonstrated superior performance with high-dimensional and highly structured data such as genomic sequencing data.
10
Reimers, Fernando M. "Conclusions. Seven Lessons to Build an Education Renaissance After the Pandemic." In Implementing Deeper Learning and 21st Education Reforms, 171–98. Cham: Springer International Publishing, 2020. http://dx.doi.org/10.1007/978-3-030-57039-2_8.
Full textAbstract:
Abstract This chapter draws out seven lessons from the cross-country analysis of the six reforms studied in this chapter. These are: Lesson 1. The power of complex mindsets about education reform. The six reforms all reflect reliance on the worldviews presented in the five frames of reform: cultural, psychological, professional, institutional and political. Those that have been sustained relied on insights from more of these five frames than those that were short lived. Lesson 2. Implementation matters considerably. The chapter discusses how the implementation process in effect recreates a reform, and how the development of an operational strategy defining the details of reform is what in the end most matters to the success of reform. The chapter discusses how the six reforms produced rather distinct operational strategies of seemingly similar components of the reform such as the learning goals for students or teacher professional development. Implementation strategies are also based on implicit theories of how organizations work, and the chapter explains the usefulness of a developmental theory of how organizations evolve to designing strategies that are aligned with the functionings that are possible in a given developmental stage, while also helping the organization evolve towards higher levels of functioning. Lesson 3. The need for operational clarity. People can’t execute what they don’t understand, and a reform must be able to translate goals into clear objectives and reform components into clear tasks which can be widely communicated and understood, as well as tracked to discern improvement and course correct when necessary. Lesson 4. Large scale reform is a journey: Coherence, Completeness and the Five Frames. The chapter explains how using the five dimensional theory of educational change can support coherence and completeness in a reform. Lesson 5. Sequencing, pacing and the importance of first steps. An operational strategy needs to be sequenced attending to ambition of goals, to existing levels of capacity and to institutional stage of development of the system. The first steps in the sequence are consequential because they shape the narrative of reform in ways that have long lasting consequences. Lesson 6. Staying the course. Long policy cycles are essential for reforms to be implemented and to produce results, and those cannot be taken for granted. Coherence, communication and participation can garner support that sustains a reform over time. Lesson 7. Learning from experience to build system level capacity. Most important to the coherent implementation of a reform is to create opportunities for key stakeholders, at various levels of the system, to learn together as a result of implementing components of the reform. Creating feedback loops and processes for making sense of such information is critical to support such learning.
Conference papers on the topic "Large dimensional learning"
1
Tiomoko, Malik, Cosme Louart, and Romain Couillet. "Large Dimensional Asymptotics of Multi-Task Learning." In ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2020. http://dx.doi.org/10.1109/icassp40776.2020.9053557.
Full text
2
Zarrouk, Tayeb, Romain Couillet, Florent Chatelain, and Nicolas Le Bihan. "Performance-Complexity Trade-Off in Large Dimensional Statistics." In 2020 IEEE 30th International Workshop on Machine Learning for Signal Processing (MLSP). IEEE, 2020. http://dx.doi.org/10.1109/mlsp49062.2020.9231568.
Full text
3
Mai, Xiaoyi, and Romain Couillet. "Revisiting and Improving Semi-supervised Learning: A Large Dimensional Approach." In ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2019. http://dx.doi.org/10.1109/icassp.2019.8683378.
Full text
4
Couillet, Romain, and Matthew McKay. "Robust covariance estimation and linear shrinkage in the large dimensional regime." In 2014 IEEE 24th International Workshop on Machine Learning for Signal Processing (MLSP). IEEE, 2014. http://dx.doi.org/10.1109/mlsp.2014.6958867.
Full text
5
Tsymbal, Alexey, Sonja Zillner, and Martin Huber. "Feature Ontology for Improved Learning from Large-Dimensional Disease-Specific Heterogeneous Data." In Twentieth IEEE International Symposium on Computer-Based Medical Systems. IEEE, 2007. http://dx.doi.org/10.1109/cbms.2007.50.
Full text
6
Liao, Zhenyu, and Romain Couillet. "Random matrices meet machine learning: A large dimensional analysis of LS-SVM." In 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2017. http://dx.doi.org/10.1109/icassp.2017.7952586.
Full text
7
Couillet, Romain. "A Random Matrix Analysis and Optimization Framework to Large Dimensional Transfer Learning." In 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP). IEEE, 2019. http://dx.doi.org/10.1109/camsap45676.2019.9022482.
Full text
8
Yu, Wenjian, Yu Gu, Jian Li, Shenghua Liu, and Yaohang Li. "Single-Pass PCA of Large High-Dimensional Data." In Twenty-Sixth International Joint Conference on Artificial Intelligence. California: International Joint Conferences on Artificial Intelligence Organization, 2017. http://dx.doi.org/10.24963/ijcai.2017/468.
Full textAbstract:
Principal component analysis (PCA) is a fundamental dimension reduction tool in statistics and machine learning. For large and high-dimensional data, computing the PCA (i.e., the top singular vectors of the data matrix) becomes a challenging task. In this work, a single-pass randomized algorithm is proposed to compute PCA with only one pass over the data. It is suitable for processing extremely large and high-dimensional data stored in slow memory (hard disk) or the data generated in a streaming fashion. Experiments with synthetic and real data validate the algorithm's accuracy, which has orders of magnitude smaller error than an existing single-pass algorithm. For a set of high-dimensional data stored as a 150 GB file, the algorithm is able to compute the first 50 principal components in just 24 minutes on a typical 24-core computer, with less than 1 GB memory cost.
9
Hongqing, Zhang, Wan Wuyi, Bao Zhongjin, Hu Jinchun, and Ye Long. "Three-dimensional Numerical Simulation for a Spillway Tunnel with High Head and Large Discharge." In The 1st EAI International Conference on Multimedia Technology and Enhanced Learning. EAI, 2017. http://dx.doi.org/10.4108/eai.28-2-2017.152332.
Full text
10
Wang, Haobo, Weiwei Liu, Yang Zhao, Tianlei Hu, Ke Chen, and Gang Chen. "Learning From Multi-Dimensional Partial Labels." In Twenty-Ninth International Joint Conference on Artificial Intelligence and Seventeenth Pacific Rim International Conference on Artificial Intelligence {IJCAI-PRICAI-20}. California: International Joint Conferences on Artificial Intelligence Organization, 2020. http://dx.doi.org/10.24963/ijcai.2020/407.
Full textAbstract:
Multi-dimensional classification has attracted huge attention from the community. Though most studies consider fully annotated data, in real practice obtaining fully labeled data in MDC tasks is usually intractable. In this paper, we propose a novel learning paradigm: MultiDimensional Partial Label Learning (MDPL) where the ground-truth labels of each instance are concealed in multiple candidate label sets. We first introduce the partial hamming loss for MDPL that incurs a large loss if the predicted labels are not in candidate label sets, and provide an empirical risk minimization (ERM) framework. Theoretically, we rigorously prove the conditions for ERM learnability of MDPL in both independent and dependent cases. Furthermore, we present two MDPL algorithms under our proposed ERM framework. Comprehensive experiments on both synthetic and real-world datasets validate the effectiveness of our proposals.
Reports on the topic "Large dimensional learning"
1
Bednar, Amy. Topological data analysis : an overview. Engineer Research and Development Center (U.S.), June 2021. http://dx.doi.org/10.21079/11681/40943.
Full textAbstract:
A growing area of mathematics topological data analysis (TDA) uses fundamental concepts of topology to analyze complex, high-dimensional data. A topological network represents the data, and the TDA uses the network to analyze the shape of the data and identify features in the network that correspond to patterns in the data. These patterns extract knowledge from the data. TDA provides a framework to advance machine learning’s ability to understand and analyze large, complex data. This paper provides background information about TDA, TDA applications for large data sets, and details related to the investigation and implementation of existing tools and environments.
2
Pritchett, Lant, and Martina Viarengo. Learning Outcomes in Developing Countries: Four Hard Lessons from PISA-D. Research on Improving Systems of Education (RISE), April 2021. http://dx.doi.org/10.35489/bsg-rise-wp_2021/069.
Full textAbstract:
The learning crisis in developing countries is increasingly acknowledged (World Bank, 2018). The UN’s Sustainable Development Goals (SDG) include goals and targets for universal learning and the World Bank has adopted a goal of eliminating learning poverty. We use student level PISA-D results for seven countries (Cambodia, Ecuador, Guatemala, Honduras, Paraguay, Senegal, and Zambia) to examine inequality in learning outcomes at the global, country, and student level for public school students. We examine learning inequality using five dimensions of potential social disadvantage measured in PISA: sex, rurality, home language, immigrant status, and socio-economic status (SES)—using the PISA measure of ESCS (Economic, Social, and Cultural Status) to measure SES. We document four important facts. First, with the exception of Ecuador, less than a third of the advantaged (male, urban, native, home speakers of the language of instruction) and ESCS elite (plus 2 standard deviations above the mean) children enrolled in public schools in PISA-D countries reach the SDG minimal target of PISA level 2 or higher in mathematics (with similarly low levels for reading and science). Even if learning differentials of enrolled students along all five dimensions of disadvantage were eliminated, the vast majority of children in these countries would not reach the SDG minimum targets. Second, the inequality in learning outcomes of the in-school children who were assessed by the PISA by household ESCS is mostly smaller in these less developed countries than in OECD or high-performing non-OECD countries. If the PISA-D countries had the same relationship of learning to ESCS as Denmark (as an example of a typical OECD country) or Vietnam (a high-performing developing country) their enrolled ESCS disadvantaged children would do worse, not better, than they actually do. Third, the disadvantages in learning outcomes along four characteristics: sex, rurality, home language, and being an immigrant country are absolutely large, but still small compared to the enormous gap between the advantaged, ESCS average students, and the SDG minimums. Given the massive global inequalities, remediating within-country inequalities in learning, while undoubtedly important for equity and justice, leads to only modest gains towards the SDG targets. Fourth, even including both public and private school students, there are strikingly few children in PISA-D countries at high levels of performance. The absolute number of children at PISA level 4 or above (reached by roughly 30 percent of OECD children) in the low performing PISA-D countries is less than a few thousand individuals, sometimes only a few hundred—in some subjects and countries just double or single digits. These four hard lessons from PISA-D reinforce the need to address global equity by “raising the floor” and targeting low learning levels (Crouch and Rolleston, 2017; Crouch, Rolleston, and Gustafsson, 2020). As Vietnam and other recent successes show, this can be done in developing country settings if education systems align around learning to improve the effectiveness of the teaching and learning processes to improve early learning of foundational skills.
3
McKenna, Patrick, and Mark Evans. Emergency Relief and complex service delivery: Towards better outcomes. Queensland University of Technology, June 2021. http://dx.doi.org/10.5204/rep.eprints.211133.
Full textAbstract:
Emergency Relief (ER) is a Department of Social Services (DSS) funded program, delivered by 197 community organisations (ER Providers) across Australia, to assist people facing a financial crisis with financial/material aid and referrals to other support programs. ER has been playing this important role in Australian communities since 1979. Without ER, more people living in Australia who experience a financial crisis might face further harm such as crippling debt or homelessness. The Emergency Relief National Coordination Group (NCG) was established in April 2020 at the start of the COVID-19 pandemic to advise the Minister for Families and Social Services on the implementation of ER. To inform its advice to the Minister, the NCG partnered with the Institute for Governance at the University of Canberra to conduct research to understand the issues and challenges faced by ER Providers and Service Users in local contexts across Australia. The research involved a desktop review of the existing literature on ER service provision, a large survey which all Commonwealth ER Providers were invited to participate in (and 122 responses were received), interviews with a purposive sample of 18 ER Providers, and the development of a program logic and theory of change for the Commonwealth ER program to assess progress. The surveys and interviews focussed on ER Provider perceptions of the strengths, weaknesses, future challenges, and areas of improvement for current ER provision. The trend of increasing case complexity, the effectiveness of ER service delivery models in achieving outcomes for Service Users, and the significance of volunteering in the sector were investigated. Separately, an evaluation of the performance of the NCG was conducted and a summary of the evaluation is provided as an appendix to this report. Several themes emerged from the review of the existing literature such as service delivery shortcomings in dealing with case complexity, the effectiveness of case management, and repeat requests for service. Interviews with ER workers and Service Users found that an uplift in workforce capability was required to deal with increasing case complexity, leading to recommendations for more training and service standards. Several service evaluations found that ER delivered with case management led to high Service User satisfaction, played an integral role in transforming the lives of people with complex needs, and lowered repeat requests for service. A large longitudinal quantitative study revealed that more time spent with participants substantially decreased the number of repeat requests for service; and, given that repeat requests for service can be an indicator of entrenched poverty, not accessing further services is likely to suggest improvement. The interviews identified the main strengths of ER to be the rapid response and flexible use of funds to stabilise crisis situations and connect people to other supports through strong local networks. Service Users trusted the system because of these strengths, and ER was often an access point to holistic support. There were three main weaknesses identified. First, funding contracts were too short and did not cover the full costs of the program—in particular, case management for complex cases. Second, many Service Users were dependent on ER which was inconsistent with the definition and intent of the program. Third, there was inconsistency in the level of service received by Service Users in different geographic locations. These weaknesses can be improved upon with a joined-up approach featuring co-design and collaborative governance, leading to the successful commissioning of social services. The survey confirmed that volunteers were significant for ER, making up 92% of all workers and 51% of all hours worked in respondent ER programs. Of the 122 respondents, volunteers amounted to 554 full-time equivalents, a contribution valued at $39.4 million. In total there were 8,316 volunteers working in the 122 respondent ER programs. The sector can support and upskill these volunteers (and employees in addition) by developing scalable training solutions such as online training modules, updating ER service standards, and engaging in collaborative learning arrangements where large and small ER Providers share resources. More engagement with peak bodies such as Volunteering Australia might also assist the sector to improve the focus on volunteer engagement. Integrated services achieve better outcomes for complex ER cases—97% of survey respondents either agreed or strongly agreed this was the case. The research identified the dimensions of service integration most relevant to ER Providers to be case management, referrals, the breadth of services offered internally, co-location with interrelated service providers, an established network of support, workforce capability, and Service User engagement. Providers can individually focus on increasing the level of service integration for their ER program to improve their ability to deal with complex cases, which are clearly on the rise. At the system level, a more joined-up approach can also improve service integration across Australia. The key dimensions of this finding are discussed next in more detail. Case management is key for achieving Service User outcomes for complex cases—89% of survey respondents either agreed or strongly agreed this was the case. Interviewees most frequently said they would provide more case management if they could change their service model. Case management allows for more time spent with the Service User, follow up with referral partners, and a higher level of expertise in service delivery to support complex cases. Of course, it is a costly model and not currently funded for all Service Users through ER. Where case management is not available as part of ER, it might be available through a related service that is part of a network of support. Where possible, ER Providers should facilitate access to case management for Service Users who would benefit. At a system level, ER models with a greater component of case management could be implemented as test cases. Referral systems are also key for achieving Service User outcomes, which is reflected in the ER Program Logic presented on page 31. The survey and interview data show that referrals within an integrated service (internal) or in a service hub (co-located) are most effective. Where this is not possible, warm referrals within a trusted network of support are more effective than cold referrals leading to higher take-up and beneficial Service User outcomes. However, cold referrals are most common, pointing to a weakness in ER referral systems. This is because ER Providers do not operate or co-locate with interrelated services in many cases, nor do they have the case management capacity to provide warm referrals in many other cases. For mental illness support, which interviewees identified as one of the most difficult issues to deal with, ER Providers offer an integrated service only 23% of the time, warm referrals 34% of the time, and cold referrals 43% of the time. A focus on referral systems at the individual ER Provider level, and system level through a joined-up approach, might lead to better outcomes for Service Users. The program logic and theory of change for ER have been documented with input from the research findings and included in Section 4.3 on page 31. These show that ER helps people facing a financial crisis to meet their immediate needs, avoid further harm, and access a path to recovery. The research demonstrates that ER is fundamental to supporting vulnerable people in Australia and should therefore continue to be funded by government.