2017 年的期刊文章 · 第 31 页

共 1233 篇,按发表时间排,新的在前,每页 40 篇。这一页是 2017 到 2017。

按期刊、按主题看 →

第 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 页

JMLR Bayesian Methods · 发表于 2017

Averaged Collapsed Variational Bayes Inference

Katsuhiko Ishiguro, Issei Sato, Naonori Ueda

摘要

This paper presents the Averaged CVB (ACVB) inference and offers convergence-guaranteed and practically useful fast Collapsed Variational Bayes (CVB) inferences. CVB inferences yield more precise inferences of Bayesian probabilistic models than Variational Bayes (VB) inferences. However, their convergence aspect is fairly unknown and has not been scrutinized. To make CVB more useful, we study their convergence behaviors in a empirical and practical approach. We develop a convergence- guaranteed algorithm for any CVB-based inference called ACVB, which enables automatic convergence detection and frees non- expert practitioners from the difficult and costly manual monitoring of inference processes. In experiments, ACVB inferences are comparable to or better than those of existing inference methods and deterministic, fast, and provide easier convergence detection. These features are especially convenient for practitioners who want precise Bayesian inference with assured convergence.

JMLR Bayesian Methods · 发表于 2017

Automatic Differentiation Variational Inference

Alp Kucukelbir, Dustin Tran, Rajesh Ranganath, Andrew Gelman, David M. Blei

摘要

Probabilistic modeling is iterative. A scientist posits a simple model, fits it to her data, refines it according to her analysis, and repeats. However, fitting complex models to large data is a bottleneck in this process. Deriving algorithms for new models can be both mathematically and computationally challenging, which makes it difficult to efficiently cycle through the steps. To this end, we develop ADVI. Using our method, the scientist only provides a probabilistic model and a dataset, nothing else. ADVI automatically derives an efficient variational inference algorithm, freeing the scientist to refine and explore many models. ADVI supports a broad class of models ---no conjugacy assumptions are required. We study ADVI across ten modern probabilistic models and apply it to a dataset with millions of observations. We deploy ADVI as part of Stan, a probabilistic programming system.

JMLR Optimization & Learning Theory · 发表于 2017

Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA

Lars Kotthoff, Chris Thornton, Holger H. Hoos, Frank Hutter, Kevin Leyton-Brown

摘要

WEKA is a widely used, open-source machine learning platform. Due to its intuitive interface, it is particularly popular with novice users. However, such users often find it hard to identify the best approach for their particular dataset among the many available. We describe the new version of Auto-WEKA, a system designed to help such users by automatically searching through the joint space of WEKA's learning algorithms and their respective hyperparameter settings to maximize performance, using a state-of-the-art Bayesian optimization method. Our new package is tightly integrated with WEKA, making it just as accessible to end users as any other learning algorithm.

JMLR Other · 发表于 2017

Asymptotic behavior of Support Vector Machine for spiked population model

Hanwen Huang

摘要

For spiked population model, we investigate the large dimension $N$ and large sample size $M$ asymptotic behavior of the Support Vector Machine (SVM) classification method in the limit of $N,M\rightarrow\infty$ at fixed $\alpha=M/N$. We focus on the generalization performance by analytically evaluating the angle between the normal direction vectors of SVM separating hyperplane and corresponding Bayes optimal separating hyperplane. This is an analogous result to the one shown in Paul (2007) and Nadler (2008) for the angle between the sample eigenvector and the population eigenvector in random matrix theorem. We provide not just bound, but sharp prediction of the asymptotic behavior of SVM that can be determined by a set of nonlinear equations. Based on the analytical results, we propose a new method of selecting tuning parameter which significantly reduces the computational cost. A surprising finding is that SVM achieves its best performance at small value of the tuning parameter under spiked population model. These results are confirmed to be correct by comparing with those of numerical simulations on finite-size systems. We also apply our formulas to an actual dataset of breast cancer and find agreement between analytical derivations and numerical computations based on cross validation.

JMLR Other · 发表于 2017

Asymptotic Analysis of Objectives Based on Fisher Information in Active Learning

Jamshid Sourati, Murat Akcakaya, Todd K. Leen, Deniz Erdogmus, Jennifer G. Dy

摘要

Obtaining labels can be costly and time-consuming. Active learning allows a learning algorithm to intelligently query samples to be labeled for a more efficient learning. Fisher information ratio (FIR) has been used as an objective for selecting queries. However, little is known about the theory behind the use of FIR for active learning. There is a gap between the underlying theory and the motivation of its usage in practice. In this paper, we attempt to fill this gap and provide a rigorous framework for analyzing existing FIR-based active learning methods. In particular, we show that FIR can be asymptotically viewed as an upper bound of the expected variance of the log-likelihood ratio. Additionally, our analysis suggests a unifying framework that not only enables us to make theoretical comparisons among the existing querying methods based on FIR, but also allows us to give insight into the development of new active learning approaches based on this objective.

Econometrica Econometrics & Economics · 发表于 2017

Assessment of Uncertainty in High Frequency Data: The Observed Asymptotic Variance

Per A. Mykland, Lan Zhang

摘要

The availability of high frequency financial data has generated a series of estimators based on intra-day data, improving the quality of large areas of financial econometrics. However, estimating the standard error of these estimators is often challenging. The root of the problem is that traditionally, standard errors rely on estimating a theoretically derived asymptotic variance, and often this asymptotic variance involves substantially more complex quantities than the original parameter to be estimated. Standard errors are important: they are used both to assess the precision of estimators in the form of confidence intervals, to create “feasible statistics” for testing, to build forecasting models based on, say, daily estimates, and also to optimize the tuning parameters. The contribution of this paper is to provide an alternative and general solution to this problem, which we call Observed Asymptotic Variance. It is a general nonparametric method for assessing asymptotic variance (AVAR). It provides consistent estimators of AVAR for a broad class of integrated parameters Θ =\int θ_t dt, where the spot parameter process θ can be a general semi-martingale, with continuous and jump components. The observed AVAR is implemented with the help of a two-scales method. Its construction works well in the presence of microstructure noise, and when the observation times are irregular or asynchronous in the multivariate case. The methodology is valid for a wide variety of estimators, including the standard ones for variance and covariance, and also for more complex estimators, such as, of leverage effects, high frequency betas, and semi-variance. *(abstract via OpenAlex)*

Econometrica Econometrics & Economics · 发表于 2017

Aspirations and Inequality

Garance Genicot, Debraj Ray

摘要

This paper develops a theory of socially determined aspirations, and the interaction of those aspirations with growth and inequality.The interaction is bidirectional: economy-wide outcomes determine individual aspirations, which in turn determine investment incentives and social outcomes.Thus aspirations, income and the distribution of income evolve jointly.When capital stocks lie in some compact set, steady states distributions must exhibit inequality and are typically clustered around local poles.When sustained growth is possible, initial histories matter.Either there is convergence to an equal distribution (with growth) or there is perennial relative divergence across clusters, with withincluster convergence.A central feature that drives these results is that aspirations that are moderately above an individual's current standard of living tend to encourage investment, while still higher aspirations may lead to frustration. *(abstract via OpenAlex)*

JMLR High-Dimensional Nonparametric & Semiparametric · 发表于 2017

Approximation Vector Machines for Large-scale Online Learning

Trung Le, Tu Dinh Nguyen, Vu Nguyen, Dinh Phung

摘要

One of the most challenging problems in kernel online learning is to bound the model size and to promote model sparsity. Sparse models not only improve computation and memory usage, but also enhance the generalization capacity -- a principle that concurs with the law of parsimony. However, inappropriate sparsity modeling may also significantly degrade the performance. In this paper, we propose Approximation Vector Machine (AVM), a model that can simultaneously encourage sparsity and safeguard its risk in compromising the performance. In an online setting context, when an incoming instance arrives, we approximate this instance by one of its neighbors whose distance to it is less than a predefined threshold. Our key intuition is that since the newly seen instance is expressed by its nearby neighbor the optimal performance can be analytically formulated and maintained. We develop theoretical foundations to support this intuition and further establish an analysis for the common loss functions including Hinge, smooth Hinge, and Logistic (i.e., for the classification task) and $\ell_{1}$, $\ell_{2}$, and $\varepsilon$-insensitive (i.e., for the regression task) to characterize the gap between the approximation and optimal solutions. This gap crucially depends on two key factors including the frequency of approximation (i.e., how frequent the approximation operation takes place) and the predefined threshold. We conducted extensive experiments for classification and regression tasks in batch and online modes using several benchmark datasets. The quantitative results show that our proposed AVM obtained comparable predictive performances with current state-of-the-art methods while simultaneously achieving significant computational speed-up due to the ability of the proposed AVM in maintaining the model size.

JMLR Other · 发表于 2017

Angle-based Multicategory Distance-weighted SVM

Hui Sun, Bruce A. Craig, Lingsong Zhang

摘要

Classification is an important supervised learning technique with numerous applications. We develop an angle-based multicategory distance-weighted support vector machine (MDWSVM) classification method that is motivated from the binary distance-weighted support vector machine (DWSVM) classification method. The new method has the merits of both support vector machine (SVM) and distance-weighted discrimination (DWD) but also alleviates both the data piling issue of SVM and the imbalanced data issue of DWD. Theoretical and numerical studies demonstrate the advantages of MDWSVM method over existing angle-based methods.

JMLR Latent Variables & Mixtures · 发表于 2017

Analyzing Tensor Power Method Dynamics in Overcomplete Regime

Animashree Anandkumar, Rong Ge, Majid Janzamin

摘要

We present a novel analysis of the dynamics of tensor power iterations in the overcomplete regime where the tensor CP rank is larger than the input dimension. Finding the CP decomposition of an overcomplete tensor is NP-hard in general. We consider the case where the tensor components are randomly drawn, and show that the simple power iteration recovers the components with bounded error under mild initialization conditions. We apply our analysis to unsupervised learning of latent variable models, such as multi-view mixture models and spherical Gaussian mixtures. Given the third order moment tensor, we learn the parameters using tensor power iterations. We prove it can correctly learn the model parameters when the number of hidden components $k$ is much larger than the data dimension $d$, up to $k = o(d^{1.5})$. We initialize the power iterations with data samples and prove its success under mild conditions on the signal-to-noise ratio of the samples. Our analysis significantly expands the class of latent variable models where spectral methods are applicable. Our analysis also deals with noise in the input tensor leading to sample complexity result in the application to learning latent variable models.

JMLR Optimization & Learning Theory · 发表于 2017

An Optimal Algorithm for Bandit and Zero-Order Convex Optimization with Two-Point Feedback

Ohad Shamir

摘要

We consider the closely related problems of bandit convex optimization with two-point feedback, and zero-order stochastic convex optimization with two function evaluations per round. We provide a simple algorithm and analysis which is optimal for convex Lipschitz functions. This improves on Duchi et al. (2015), which only provides an optimal result for smooth functions; Moreover, the algorithm and analysis are simpler, and readily extend to non-Euclidean problems. The algorithm is based on a small but surprisingly powerful modification of the gradient estimator.

JMLR Other · 发表于 2017

An Easy-to-hard Learning Paradigm for Multiple Classes and Multiple Labels

Weiwei Liu, Ivor W. Tsang, Klaus-Robert Müller

摘要

Many applications, such as human action recognition and object detection, can be formulated as a multiclass classification problem. One-vs-rest (OVR) is one of the most widely used approaches for multiclass classification due to its simplicity and excellent performance. However, many confusing classes in such applications will degrade its results. For example, hand clap and boxing are two confusing actions. Hand clap is easily misclassified as boxing, and vice versa. Therefore, precisely classifying confusing classes remains a challenging task. To obtain better performance for multiclass classifications that have confusing classes, we first develop a classifier chain model for multiclass classification (CCMC) to transfer class information between classifiers. Then, based on an analysis of our proposed model, we propose an easy- to-hard learning paradigm for multiclass classification to automatically identify easy and hard classes and then use the predictions from simpler classes to help solve harder classes. Similar to CCMC, the classifier chain (CC) model is also proposed by Read et al. (2009) to capture the label dependency for multi-label classification. However, CC does not consider the order of difficulty of the labels and achieves degenerated performance when there are many confusing labels. Therefore, it is non- trivial to learn the appropriate label order for CC. Motivated by our analysis for CCMC, we also propose the easy-to-hard learning paradigm for multi-label classification to automatically identify easy and hard labels, and then use the predictions from simpler labels to help solve harder labels. We also demonstrate that our proposed strategy can be successfully applied to a wide range of applications, such as ordinal classification and relationship prediction. Extensive empirical studies validate our analysis and the effectiveness of our proposed easy-to-hard learning strategies.

Econometrica Networks & Graphs Econometrics & Economics · 发表于 2017

Altruism in Networks

Renaud Bourlès, Yann Bramoullé, Eduardo Perez-Richet

摘要

We provide the first analysis of altruism in networks. Agents are connected through a fixed, weighted network and care about the well-being of their network neigh-bors. Given some initial distribution of incomes, agents may provide financial support to their poorer friends. We characterize the Nash equilibria of this transfer game for general networks and utility functions. We show that equilibria solve a well-behaved maximization program, related to classical problems of optimal transportation on networks. We build on this reformulation and establish existence, uniqueness in consumptions and generic uniqueness in transfers. We show that transfers are affected by the geometry of the altru-istic network. They flow through shortest paths and chains of transfers emerge when the network is not transitive. We analyze the effects of changes in incomes and in the network. When an agent suffers a negative income shock, the equilibrium consumption of every agent decreases weakly. We characterize the impact of small redistributions and show that decreasing income inequality may increase consumption inequality. We also characterize the impact of a small increase in altruism. While altruistic networks reduce inequality, more altruism may lead to more inequality. *(abstract via OpenAlex)*

JMLR Other · 发表于 2017

Adaptive Randomized Dimension Reduction on Massive Data

Gregory Darnell, Stoyan Georgiev, Sayan Mukherjee, Barbara E Engelhardt

摘要

The scalability of statistical estimators is of increasing importance in modern applications. One approach to implementing scalable algorithms is to compress data into a low dimensional latent space using dimension reduction methods. In this paper, we develop an approach for dimension reduction that exploits the assumption of low rank structure in high dimensional data to gain both computational and statistical advantages. We adapt recent randomized low-rank approximation algorithms to provide an efficient solution to principal component analysis (PCA), and we use this efficient solver to improve estimation in large- scale linear mixed models (LMM) for association mapping in statistical genomics. A key observation in this paper is that randomization serves a dual role, improving both computational and statistical performance by implicitly regularizing the covariance matrix estimate of the random effect in an LMM. These statistical and computational advantages are highlighted in our experiments on simulated data and large-scale genomic studies.

JMLR Other · 发表于 2017

Active-set Methods for Submodular Minimization Problems

K. S. Sesh Kumar, Francis Bach

摘要

We consider the submodular function minimization (SFM) and the quadratic minimization problems regularized by the Lovasz extension of the submodular function. These optimization problems are intimately related; for example, min-cut problems and total variation denoising problems, where the cut function is submodular and its Lovasz extension is given by the associated total variation. When a quadratic loss is regularized by the total variation of a cut function, it thus becomes a total variation denoising problem and we use the same terminology in this paper for general submodular functions. We propose a new active-set algorithm for total variation denoising with the assumption of an oracle that solves the corresponding SFM problem. This can be seen as local descent algorithm over ordered partitions with explicit convergence guarantees. It is more flexible than the existing algorithms with the ability for warm-restarts using the solution of a closely related problem. Further, we also consider the case when a submodular function can be decomposed into the sum of two submodular functions $F_1$ and $F_2$ and assume SFM oracles for these two functions. We propose a new active-set algorithm for total variation denoising (and hence SFM by thresholding the solution at zero). This algorithm also performs local descent over ordered partitions and its ability to warm start considerably improves the performance of the algorithm. In the experiments, we compare the performance of the proposed algorithms with state-of-the-art algorithms, showing that it reduces the calls to SFM oracles.

JMLR Missing Data & Measurement Error · 发表于 2017

Achieving Optimal Misclassification Proportion in Stochastic Block Models

Chao Gao, Zongming Ma, Anderson Y. Zhang, Harrison H. Zhou

摘要

Community detection is a fundamental statistical problem in network data analysis. In this paper, we present a polynomial time two-stage method that provably achieves optimal statistical performance in misclassification proportion for stochastic block model under weak regularity conditions. Our two-stage procedure consists of a refinement stage motivated by penalized local maximum likelihood estimation. This stage can take a wide range of weakly consistent community detection procedures as its initializer, to which it applies and outputs a community assignment that achieves optimal misclassification proportion with high probability. The theoretical property is confirmed by simulated examples.

SiM Survival Analysis Clinical Trials · 发表于 2017

Accounting for competing risks in randomized controlled trials: a review and recommendations for improvement

Peter C. Austin, Jason P. Fine

摘要

In studies with survival or time‐to‐event outcomes, a competing risk is an event whose occurrence precludes the occurrence of the primary event of interest. Specialized statistical methods must be used to analyze survival data in the presence of competing risks. We conducted a review of randomized controlled trials with survival outcomes that were published in high‐impact general medical journals. Of 40 studies that we identified, 31 (77.5%) were potentially susceptible to competing risks. However, in the majority of these studies, the potential presence of competing risks was not accounted for in the statistical analyses that were described. Of the 31 studies potentially susceptible to competing risks, 24 (77.4%) reported the results of a Kaplan–Meier survival analysis, while only five (16.1%) reported using cumulative incidence functions to estimate the incidence of the outcome over time in the presence of competing risks. The former approach will tend to result in an overestimate of the incidence of the outcome over time, while the latter approach will result in unbiased estimation of the incidence of the primary outcome over time. We provide recommendations on the analysis and reporting of randomized controlled trials with survival outcomes in the presence of competing risks. © 2017 The Authors. Statistics in Medicine

JMLR Optimization & Learning Theory · 发表于 2017

Accelerating Stochastic Composition Optimization

Mengdi Wang, Ji Liu, Ethan X. Fang

摘要

We consider the stochastic nested composition optimization problem where the objective is a composition of two expected- value functions. We propose a new stochastic first-order method, namely the accelerated stochastic compositional proximal gradient (ASC-PG) method. This algorithm updates the solution based on noisy gradient queries using a two-timescale iteration. The ASC-PG is the first proximal gradient method for the stochastic composition problem that can deal with nonsmooth regularization penalty. We show that the ASC-PG exhibits faster convergence than the best known algorithms, and that it achieves the optimal sample-error complexity in several important special cases. We demonstrate the application of ASC-PG to reinforcement learning and conduct numerical experiments.

JMLR Survey & Sampling Optimization & Learning Theory · 发表于 2017

A survey of Algorithms and Analysis for Adaptive Online Learning

H. Brendan McMahan

摘要

We present tools for the analysis of Follow-The-Regularized- Leader (FTRL), Dual Averaging, and Mirror Descent algorithms when the regularizer (equivalently, prox-function or learning rate schedule) is chosen adaptively based on the data. Adaptivity can be used to prove regret bounds that hold on every round, and also allows for data-dependent regret bounds as in AdaGrad-style algorithms (e.g., Online Gradient Descent with adaptive per-coordinate learning rates). We present results from a large number of prior works in a unified manner, using a modular and tight analysis that isolates the key arguments in easily re-usable lemmas. This approach strengthens previously known FTRL analysis techniques to produce bounds as tight as those achieved by potential functions or primal-dual analysis. Further, we prove a general and exact equivalence between adaptive Mirror Descent algorithms and a corresponding FTRL update, which allows us to analyze Mirror Descent algorithms in the same framework. The key to bridging the gap between Dural Averaging and Mirror Descent algorithms lies in an analysis of the FTRL-Proximal algorithm family. Our regret bounds are proved in the most general form, holding for arbitrary norms and non- smooth regularizers with time-varying weight.

SiM Causal Inference Genetics & Omics · 发表于 2017

A framework for the investigation of pleiotropy in two‐sample summary data Mendelian randomization

Jack Bowden, Fabiola Del Greco M, Cosetta Minelli, George Davey Smith, Nuala Sheehan, John Thompson

摘要

Mendelian randomization (MR) uses genetic data to probe questions of causality in epidemiological research, by invoking the Instrumental Variable (IV) assumptions. In recent years, it has become commonplace to attempt MR analyses by synthesising summary data estimates of genetic association gleaned from large and independent study populations. This is referred to as two‐sample summary data MR. Unfortunately, due to the sheer number of variants that can be easily included into summary data MR analyses, it is increasingly likely that some do not meet the IV assumptions due to pleiotropy. There is a pressing need to develop methods that can both detect and correct for pleiotropy, in order to preserve the validity of the MR approach in this context. In this paper, we aim to clarify how established methods of meta‐regression and random effects modelling from mainstream meta‐analysis are being adapted to perform this task. Specifically, we focus on two contrastin g approaches: the Inverse Variance Weighted (IVW) method which assumes in its simplest form that all genetic variants are valid IVs, and the method of MR‐Egger regression that allows all variants to violate the IV assumptions, albeit in a specific way. We investigate the ability of two popular random effects models to provide robustness to pleiotropy under the IVW approach, and propose statistics to quantify the relative goodness‐of‐fit of the IVW approach over MR‐Egger regression. © 2017 The Authors. Statistics in Medicine

JMLR Meta-Analysis & Data Integration · 发表于 2017

A distributed block coordinate descent method for training l1 regularized linear classifiers

Dhruv Mahajan, S. Sathiya Keerthi, S. Sundararajan

摘要

Distributed training of $l_1$ regularized classifiers has received great attention recently. Most existing methods approach this problem by taking steps obtained from approximating the objective by a quadratic approximation that is decoupled at the individual variable level. These methods are designed for multicore systems where communication costs are low. They are inefficient on systems such as Hadoop running on a cluster of commodity machines where communication costs are substantial. In this paper we design a distributed algorithm for $l_1$ regularization that is much better suited for such systems than existing algorithms. A careful cost analysis is used to support these points and motivate our method. The main idea of our algorithm is to do block optimization of many variables on the actual objective function within each computing node; this increases the computational cost per step that is matched with the communication cost, and decreases the number of outer iterations, thus yielding a faster overall method. Distributed Gauss-Seidel and Gauss-Southwell greedy schemes are used for choosing variables to update in each step. We establish global convergence theory for our algorithm, including Q-linear rate of convergence. Experiments on two benchmark problems show our method to be much faster than existing methods.

JMLR Other · 发表于 2017

A Unifying Framework for Gaussian Process Pseudo-Point Approximations using Power Expectation Propagation

Thang D. Bui, Josiah Yan, Richard E. Turner

摘要

Gaussian processes (GPs) are flexible distributions over functions that enable high-level assumptions about unknown functions to be encoded in a parsimonious, flexible and general way. Although elegant, the application of GPs is limited by computational and analytical intractabilities that arise when data are sufficiently numerous or when employing non-Gaussian models. Consequently, a wealth of GP approximation schemes have been developed over the last 15 years to address these key limitations. Many of these schemes employ a small set of pseudo data points to summarise the actual data. In this paper we develop a new pseudo-point approximation framework using Power Expectation Propagation (Power EP) that unifies a large number of these pseudo-point approximations. Unlike much of the previous venerable work in this area, the new framework is built on standard methods for approximate inference (variational free- energy, EP and Power EP methods) rather than employing approximations to the probabilistic generative model itself. In this way all of the approximation is performed at `inference time' rather than at `modelling time', resolving awkward philosophical and empirical questions that trouble previous approaches. Crucially, we demonstrate that the new framework includes new pseudo-point approximation methods that outperform current approaches on regression and classification tasks.

JMLR Optimization & Learning Theory · 发表于 2017

A Unified Formulation and Fast Accelerated Proximal Gradient Method for Classification

Naoki Ito, Akiko Takeda, Kim-Chuan Toh

摘要

Binary classification is the problem of predicting the class a given sample belongs to. To achieve a good prediction performance, it is important to find a suitable model for a given dataset. However, it is often time consuming and impractical for practitioners to try various classification models because each model employs a different formulation and algorithm. The difficulty can be mitigated if we have a unified formulation and an efficient universal algorithmic framework for various classification models to expedite the comparison of performance of different models for a given dataset. In this paper, we present a unified formulation of various classification models (including $C$-SVM, $\ell_2$-SVM, $\nu$-SVM, MM-FDA, MM-MPM, logistic regression, distance weighted discrimination) and develop a general optimization algorithm based on an accelerated proximal gradient (APG) method for the formulation. We design various techniques such as backtracking line search and adaptive restarting strategy in order to speed up the practical convergence of our method. We also give a theoretical convergence guarantee for the proposed fast APG algorithm. Numerical experiments show that our algorithm is stable and highly competitive to specialized algorithms designed for specific models (e.g., sequential minimal optimization (SMO) for SVM).

Econometrica Other · 发表于 2017

A Theory of Intergenerational Altruism

Simone Galperti, Bruno Strulovici

摘要

Modeling intergenerational altruism is crucial to evaluate the long-term consequences of current decisions, and requires a set of principles guiding such altruism. We axiomatically develop a theory of pure, direct altruism: Altruism is pure if it concerns the total utility (rather than the mere consumption utility) of future generations, and direct if it directly incorporates the utility of all future generations. Our axioms deliver a new class of altruistic, forward-looking preferences, whose weight put on the consumption of a future generation generally depends on the consumption of other generations. The only preferences lacking this dependence correspond to the quasi-hyperbolic discounting model, which our theory characterizes. Our approach provides a framework to analyze welfare in the presence of altruistic preferences and addresses technical challenges stemming from the interdependent nature of such preferences. *(abstract via OpenAlex)*

JMLR Survey & Sampling Machine Learning · 发表于 2017

A Survey of Preference-Based Reinforcement Learning Methods

Christian Wirth, Riad Akrour, Gerhard Neumann, Johannes Fürnkranz

摘要

Reinforcement learning (RL) techniques optimize the accumulated long-term reward of a suitably chosen reward function. However, designing such a reward function often requires a lot of task- specific prior knowledge. The designer needs to consider different objectives that do not only influence the learned behavior but also the learning progress. To alleviate these issues, preference-based reinforcement learning algorithms (PbRL) have been proposed that can directly learn from an expert's preferences instead of a hand-designed numeric reward. PbRL has gained traction in recent years due to its ability to resolve the reward shaping problem, its ability to learn from non numeric rewards and the possibility to reduce the dependence on expert knowledge. We provide a unified framework for PbRL that describes the task formally and points out the different design principles that affect the evaluation task for the human as well as the computational complexity. The design principles include the type of feedback that is assumed, the representation that is learned to capture the preferences, the optimization problem that has to be solved as well as how the exploration/exploitation problem is tackled. Furthermore, we point out shortcomings of current algorithms, propose open research questions and briefly survey practical tasks that have been solved using PbRL.

Econometrica Networks & Graphs · 发表于 2017

A Structural Model of Dense Network Formation

Angelo Mele

摘要

This paper proposes an empirical model of network formation, combining strategic and random networks features. Payoffs depend on direct links, but also link externalities. Players meet sequentially at random, myopically updating their links. Under mild assumptions, the network formation process is a potential game and converges to an exponential random graph model (ERGM), generating directed dense networks. I provide new identification results for ERGMs in large networks: if link externalities are nonnegative, the ERGM is asymptotically indistinguishable from an Erdős–Renyi model with independent links. We can identify the parameters only when at least one of the externalities is negative and sufficiently large. However, the standard estimation methods for ERGMs can have exponentially slow convergence, even when the model has asymptotically independent links. I thus estimate parameters using a Bayesian MCMC method. When the parameters are identifiable, I show evidence that the estimation algorithm converges in almost quadratic time. *(abstract via OpenAlex)*

JMLR Latent Variables & Mixtures · 发表于 2017

A Spectral Algorithm for Inference in Hidden semi-Markov Models

Igor Melnyk, Arindam Banerjee

摘要

Hidden semi-Markov models (HSMMs) are latent variable models which allow latent state persistence and can be viewed as a generalization of the popular hidden Markov models (HMMs). In this paper, we introduce a novel spectral algorithm to perform inference in HSMMs. Unlike expectation maximization (EM), our approach correctly estimates the probability of given observation sequence based on a set of training sequences. Our approach is based on estimating moments from the sample, whose number of dimensions depends only logarithmically on the maximum length of the hidden state persistence. Moreover, the algorithm requires only a few matrix inversions and is therefore computationally efficient. Empirical evaluations on synthetic and real data demonstrate the advantage of the algorithm over EM in terms of speed and accuracy, especially for large data sets.

JMLR High-Dimensional Nonparametric & Semiparametric · 发表于 2017

A Robust-Equitable Measure for Feature Ranking and Selection

A. Adam Ding, Jennifer G. Dy, Yi Li, Yale Chang

摘要

In many applications, not all the features used to represent data samples are important. Often only a few features are relevant for the prediction task. The choice of dependence measures often affect the final result of many feature selection methods. To select features that have complex nonlinear relationships with the response variable, the dependence measure should be equitable, a concept proposed by Reshef et al. (2011); that is, the dependence measure treats linear and nonlinear relationships equally. Recently, Kinney and Atwal (2014) gave a mathematical definition of self- equitability. In this paper, we introduce a new concept of robust-equitability and identify a robust- equitable copula dependence measure, the robust copula dependence (RCD) measure. RCD is based on the $L_1$-distance of the copula density from uniform and we show that it is equitable under both equitability definitions. We also prove theoretically that RCD is much easier to estimate than mutual information. Because of these theoretical properties, the RCD measure has the following advantages compared to existing dependence measures: it is robust to different relationship forms and robust to unequal sample sizes of different features. Experiments on both synthetic and real-world data sets confirm the theoretical analysis, and illustrate the advantage of using the dependence measure RCD for feature selection.

JMLR Optimization & Learning Theory · 发表于 2017

A Nonconvex Approach for Phase Retrieval: Reshaped Wirtinger Flow and Incremental Algorithms

Huishuai Zhang, Yingbin Liang, Yuejie Chi

摘要

We study the problem of solving a quadratic system of equations, i.e., recovering a vector signal $\boldsymbol{x}\in \mathbb{R}^n$ from its magnitude measurements $y_i=|\langle \boldsymbol{a}_i, \boldsymbol{x}\rangle|, i=1,..., m$. We develop a gradient descent algorithm (referred to as RWF for reshaped Wirtinger flow) by minimizing the quadratic loss of the magnitude measurements. Comparing with Wirtinger flow (WF) (Candes et al., 2015), the loss function of RWF is nonconvex and nonsmooth, but better resembles the least-squares loss when the phase information is also available. We show that for random Gaussian measurements, RWF enjoys linear convergence to the true signal as long as the number of measurements is $\mathcal{O}(n)$. This improves the sample complexity of WF ($\mathcal{O}(n\log n)$), and achieves the same sample complexity as truncated Wirtinger flow (TWF) (Chen and Candes, 2015), but without any sophisticated truncation in the gradient loop. Furthermore, RWF costs less computationally than WF, and runs faster numerically than both WF and TWF. We further develop an incremental (stochastic) version of RWF (IRWF) and connect it with the randomized Kaczmarz method for phase retrieval. We demonstrate that IRWF outperforms existing incremental as well as batch algorithms with experiments.

JMLR Meta-Analysis & Data Integration Optimization & Learning Theory · 发表于 2017

A General Distributed Dual Coordinate Optimization Framework for Regularized Loss Minimization

Shun Zheng, Jialei Wang, Fen Xia, Wei Xu, Tong Zhang

摘要

In modern large-scale machine learning applications, the training data are often partitioned and stored on multiple machines. It is customary to employ the data parallelism approach, where the aggregated training loss is minimized without moving data across machines. In this paper, we introduce a novel distributed dual formulation for regularized loss minimization problems that can directly handle data parallelism in the distributed setting. This formulation allows us to systematically derive dual coordinate optimization procedures, which we refer to as Distributed Alternating Dual Maximization (DADM). The framework extends earlier studies described in (Boyd et al., 2011; Ma et al., 2017; Jaggi et al., 2014; Yang, 2013) and has rigorous theoretical analyses. Moreover, with the help of the new formulation, we develop the accelerated version of DADM (Acc-DADM) by generalizing the acceleration technique from (Shalev-Shwartz and Zhang, 2014) to the distributed setting. We also provide theoretical results for the proposed accelerated version, and the new result improves previous ones (Yang, 2013; Ma et al., 2017) whose iteration complexities grow linearly on the condition number. Our empirical studies validate our theory and show that our accelerated approach significantly improves the previous state- of-the-art distributed dual coordinate optimization algorithms.

Econometrica Other · 发表于 2017

A Fairness Justification of Utilitarianism

Paolo Giovanni Piacquadio

摘要

Differences in preferences are important to explain variation in individuals' behavior. There is, however, no consensus on how to take these differences into account when evaluating policies. While prominent in the economic literature, the standard utilitarian criterion is controversial. According to some, interpersonal comparability of utilities involves value judgments with little objective basis. Others argue that social justice is primarily about the distribution of commodities assigned to individuals, rather than their subjective satisfaction or happiness. In this paper, we propose and axiomatically characterize a criterion, named opportunity-equivalent utilitarian, that addresses these claims. First, our criterion ranks social alternatives on the basis of individuals' ordinal preferences. Second, it compares individuals based on the fairness of their assignments. Opportunity-equivalent utilitarianism requires society to maximize the sum of specific indices of well-being that are cardinal, interpersonally comparable, and represent each individual's preferences. *(abstract via OpenAlex)*

JMLR Longitudinal Data Bayesian Methods · 发表于 2017

A Bayesian Mixed-Effects Model to Learn Trajectories of Changes from Repeated Manifold-Valued Observations

Jean-Baptiste Schiratti, Stéphanie Allassonnière, Olivier Colliot, Stanley Durrleman

摘要

We propose a generic Bayesian mixed-effects model to estimate the temporal progression of a biological phenomenon from observations obtained at multiple time points for a group of individuals. The progression is modeled by continuous trajectories in the space of measurements. Individual trajectories of progression result from spatiotemporal transformations of an average trajectory. These transformations allow for the quantification of changes in direction and pace at which the trajectories are followed. The framework of Riemannian geometry allows the model to be used with any kind of measurements with smooth constraints. A stochastic version of the Expectation-Maximization algorithm is used to produce maximum a posteriori estimates of the parameters. We evaluated our method using a series of neuropsychological test scores from patients with mild cognitive impairments, later diagnosed with Alzheimer's disease, and simulated evolutions of symmetric positive definite matrices. The data-driven model of impairment of cognitive functions illustrated the variability in the ordering and timing of the decline of these functions in the population. We showed that the estimated spatiotemporal transformations effectively put into correspondence significant events in the progression of individuals.

JMLR Bayesian Methods Machine Learning · 发表于 2017

A Bayesian Framework for Learning Rule Sets for Interpretable Classification

Tong Wang, Cynthia Rudin, Finale Doshi-Velez, Yimin Liu, Erica Klampfl, Perry MacNeille

摘要

We present a machine learning algorithm for building classifiers that are comprised of a small number of short rules. These are restricted disjunctive normal form models. An example of a classifier of this form is as follows: If $X$ satisfies (condition $A$ AND condition $B$) OR (condition $C$) OR $\cdots$, then $Y=1$. Models of this form have the advantage of being interpretable to human experts since they produce a set of rules that concisely describe a specific class. We present two probabilistic models with prior parameters that the user can set to encourage the model to have a desired size and shape, to conform with a domain-specific definition of interpretability. We provide a scalable MAP inference approach and develop theoretical bounds to reduce computation by iteratively pruning the search space. We apply our method (Bayesian Rule Sets -- BRS) to characterize and predict user behavior with respect to in-vehicle context-aware personalized recommender systems. Our method has a major advantage over classical associative classification methods and decision trees in that it does not greedily grow the model.

第 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 页