The emergence of Big Data has enabled new research perspectives in the discrete choice community. While the techniques to estimate Machine Learning models on a massive amount of data are well established, these have not yet been fully explored for the estimation of statistical Discrete Choice Models based on the random utility framework. In this article, we provide new ways of dealing with large datasets in the context of Discrete Choice Models. We achieve this by proposing new efficient stochastic optimization algorithms and extensively testing them alongside existing approaches. We develop these algorithms based on three main contributions: the use of a stochastic Hessian, the modification of the batch size, and a change of optimization algorithm depending on the batch size. A comprehensive experimental comparison of fifteen optimization algorithms is conducted across ten benchmark Discrete Choice Model cases. The results indicate that the HAMABS algorithm, a hybrid adaptive batch size stochastic method, is the best performing algorithm across the optimization benchmarks. This algorithm speeds up the optimization time by a factor of 23 on the largest model compared to existing algorithms used in practice. The integration of the new algorithms in Discrete Choice Models estimation software will significantly reduce the time required for model estimation and therefore enable researchers and practitioners to explore new approaches for the specification of choice models.
Norwegian University of Science and Technology（挪威科技大学）；BASF SE（巴斯夫公司）；Imperial College London（帝国理工学院）
Nonlinear model predictive control (NMPC) is an efficient approach for the control of nonlinear multivariable dynamic systems with constraints, which however requires an accurate plant model. Plant models can often be determined from first principles, parts of the model are however difficult to derive using physical laws alone. In this paper a hybrid Gaussian process (GP) first principles modeling scheme is proposed to overcome this issue, which exploits GPs to model the parts of the dynamic system that are difficult to describe using first principles. GPs not only give accurate predictions, b…
This paper studies the reduced-order or full-order, dead-beat observer problem for a class of nonlinear systems, linear in the unmeasured states. A novel hybrid observer design strategy is proposed, with the help of the notion of strong observability in finite time. The proposed methodology is applied to a batch reactor, for which a hybrid dead-beat observer is obtained in the absence of the precise measurements of the concentration variables. Moreover, the observer is used for the estimation of the frequency of a sinusoidal signal. The results show that accurate estimations can be provided ev…
We analyze the dynamics of large batch stochastic gradient descent with momentum (SGD+M) on the least squares problem when both the number of samples and dimensions are large. In this setting, we show that the dynamics of SGD+M converge to a deterministic discrete Volterra equation as dimension increases, which we analyze. We identify a stability measurement, the implicit conditioning ratio (ICR), which regulates the ability of SGD+M to accelerate the algorithm. When the batch size exceeds this ICR, SGD+M converges linearly at a rate of $\mathcal{O}(1/\sqrtκ)$, matching optimal full-batch mome…
The standard L-BFGS method relies on gradient approximations that are not dominated by noise, so that search directions are descent directions, the line search is reliable, and quasi-Newton updating yields useful quadratic models of the objective function. All of this appears to call for a full batch approach, but since small batch sizes give rise to faster algorithms with better generalization properties, L-BFGS is currently not considered an algorithm of choice for large-scale machine learning applications. One need not, however, choose between the two extremes represented by the full batch …
Laboratory for Computational and Statistical Learning（计算与统计学习实验室）；Istituto Italiano di Tecnologia（意大利理工学院）；Massachusetts Institute of Technology（麻省理工学院）；DIBRIS, Università di Genova（热那亚大学DIBRIS（信息学、生物信息学与系统生物学系））；Università di Genova（热那亚大学）
We analyze the learning properties of the stochastic gradient method when multiple passes over the data and mini-batches are allowed. We study how regularization properties are controlled by the step-size, the number of passes and the mini-batch size. In particular, we consider the square loss and show that for a universal step-size choice, the number of passes acts as a regularization parameter, and optimal finite sample bounds can be achieved by early-stopping. Moreover, we show that larger step-sizes are allowed when considering mini-batches. Our analysis is based on a unifying approach, en…
Laboratory for Information and Inference Systems (LIONS), EPFL（洛桑联邦理工学院信息与推理系统实验室（LIONS））；Laboratoire Traitement et Communication d’Information, Télécom Paris, Institut Polytechnique de Paris（巴黎理工学院巴黎电信学院信息处理与通信实验室）；Department of Electrical Engineering (ESAT-STADIUS), KU Leuven（荷语鲁汶大学电气工程系（ESAT-STADIUS））
This paper introduces a family of stochastic extragradient-type algorithms for a class of nonconvex-nonconcave problems characterized by the weak Minty variational inequality (MVI). Unlike existing results on extragradient methods in the monotone setting, employing diminishing stepsizes is no longer possible in the weak MVI setting. This has led to approaches such as increasing batch sizes per iteration which can however be prohibitively expensive. In contrast, our proposed methods involves two stepsizes and only requires one additional oracle evaluation per iteration. We show that it is possi…
Sun Yat-sen University（中山大学）；Xiamen University（厦门大学）
StochAstic Recursive grAdient algoritHm (SARAH), originally proposed for convex optimization and also proven to be effective for general nonconvex optimization, has received great attention due to its simple recursive framework for updating stochastic gradient estimates. The performance of SARAH significantly depends on the choice of step size sequence. However, SARAH and its variants often employ a best-tuned step size by mentor, which is time consuming in practice. Motivated by this gap, we proposed a variant of the Barzilai-Borwein (BB) method, referred to as the Random Barzilai-Borwein (RB…
Adaptive Batch Size Schedules for Distributed Training of Language Models with Data and Model Parallelism
Perelman School of Medicine, University of Pennsylvania（宾夕法尼亚大学佩雷尔曼医学院）；The University of Chicago Booth School of Business（芝加哥大学布斯商学院）；Northwestern University（西北大学）；University of Southern California Marshall School of Business（南加州大学马歇尔商学院）
We introduce a hybrid stochastic estimator to design stochastic gradient algorithms for solving stochastic optimization problems. Such a hybrid estimator is a convex combination of two existing biased and unbiased estimators and leads to some useful property on its variance. We limit our consideration to a hybrid SARAH-SGD for nonconvex expectation problems. However, our idea can be extended to handle a broader class of estimators in both convex and nonconvex settings. We propose a new single-loop stochastic gradient descent algorithm that can achieve $O(\max\{σ^3\varepsilon^{-1},σ\varepsilon^…
Massachusetts Institute of Technology（麻省理工学院）；Istituto Italiano di Tecnologia（意大利理工学院）；Università degli Studi di Genova（热那亚大学）
In the setting of nonparametric regression, we propose and study a combination of stochastic gradient methods with Nyström subsampling, allowing multiple passes over the data and mini-batches. Generalization error bounds for the studied algorithm are provided. Particularly, optimal learning rates are derived considering different possible choices of the step-size, the mini-batch size, the number of iterations/passes, and the subsampling level. In comparison with state-of-the-art algorithms such as the classic stochastic gradient methods and kernel ridge regression with Nyström, the studied alg…
On the equivalence of different adaptive batch size selection strategies for stochastic gradient descent methods
University of Nottingham（诺丁汉大学）；Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）；RWTH Aachen University（亚琛工业大学）；King Abdullah University of Science & Technology（阿卜杜拉国王科技大学）；Stevens Institute of Technology（史蒂文斯理工学院）
How Does Critical Batch Size Scale in Pre-training?
Harvard University（哈佛大学）；Kempner Institute at Harvard University（哈佛大学凯普勒研究所）；University of California, Berkeley（加州大学伯克利分校）；The University of Hong Kong（香港大学）；Amazon（亚马逊公司）
University of Texas at Austin（德克萨斯大学奥斯汀分校）；Facebook AI Research（Facebook人工智能研究院）
Adjusting the learning rate schedule in stochastic gradient methods is an important unresolved problem which requires tuning in practice. If certain parameters of the loss function such as smoothness or strong convexity constants are known, theoretical learning rate schedules can be applied. However, in practice, such parameters are not known, and the loss function of interest is not convex in any case. The recently proposed batch normalization reparametrization is widely adopted in most neural network architectures today because, among other advantages, it is robust to the choice of Lipschitz…
