arXiv subjects

James T. Kwok

Publications and source records attributed to James T. Kwok.

68 records · Page 4Linked to original sources

arXiv2018

Loss-aware Weight Quantization of Deep Networks

The huge size of deep networks hinders their use in small computing devices. In this paper, we consider compressing the network by weight quantization. We extend a recently proposed loss-aware weight binarization scheme to ternarization, with possibly different scaling parameters for the positive and negative weights, and m-bit (where m > 2) quantization. Experiments on feedforward and recurrent neural networks show that the proposed scheme outperforms state-of-the-art weight quantization algorithms, and is as accurate (or even more accurate) than the full-precision network.

cs.LG↗

arXiv2018

Power Law in Sparsified Deep Neural Networks

The power law has been observed in the degree distributions of many biological neural networks. Sparse deep neural networks, which learn an economical representation from the data, resemble biological neural networks in many ways. In this paper, we study if these artificial networks also exhibit properties of the power law. Experimental results on two popular deep learning models, namely, multilayer perceptrons and convolutional neural networks, are affirmative. The power law is also naturally related to preferential attachment. To study the dynamical properties of deep networks in continual learning, we propose an internal preferential attachment model to explain how the network topology evolves. Experimental results show that with the arrival of a new task, the new connections made follow this preferential attachment process.

cs.LG↗

arXiv2017

Scalable Online Convolutional Sparse Coding

Convolutional sparse coding (CSC) improves sparse coding by learning a shift-invariant dictionary from the data. However, existing CSC algorithms operate in the batch mode and are expensive, in terms of both space and time, on large datasets. In this paper, we alleviate these problems by using online learning. The key is a reformulation of the CSC objective so that convolution can be handled easily in the frequency domain and much smaller history matrices are needed. We use the alternating direction method of multipliers (ADMM) to solve the resulting optimization problem and the ADMM subproblems have efficient closed-form solutions. Theoretical analysis shows that the learned dictionary converges to a stationary point of the optimization problem. Extensive experiments show that convergence of the proposed method is much faster and its reconstruction performance is also better. Moreover, while existing CSC algorithms can only run on a small number of images, the proposed method can handle at least ten times more images.

cs.CV↗

arXiv2017

Collaborative Filtering with Social Local Models

Matrix Factorization (MF) is a very popular method for recommendation systems. It assumes that the underneath rating matrix is low-rank. However, this assumption can be too restrictive to capture complex relationships and interactions among users and items. Recently, Local LOw-Rank Matrix Approximation (LLORMA) has been shown to be very successful in addressing this issue. It just assumes the rating matrix is composed of a number of low-rank submatrices constructed from subsets of similar users and items. Although LLORMA outperforms MF, how to construct such submatrices remains a big problem. Motivated by the availability of rich social connections in today's recommendation systems, we propose a novel framework, i.e., Social LOcal low-rank Matrix Approximation (SLOMA), to address this problem. To the best of our knowledge, SLOMA is the first work to incorporate social connections into the local low-rank framework. Furthermore, we enhance SLOMA by applying social regularization to submatrices factorization, denoted as SLOMA++. Therefore, the proposed model can benefit from both social recommendation and the local low-rank assumption. Experimental results from two real-world datasets, Yelp and Douban, demonstrate the superiority of the proposed models over LLORMA and MF.

cs.SI↗

arXiv2017

Fast Learning with Nonconvex L1-2 Regularization

Convex regularizers are often used for sparse learning. They are easy to optimize, but can lead to inferior prediction performance. The difference of $\ell_1$ and $\ell_2$ ($\ell_{1-2}$) regularizer has been recently proposed as a nonconvex regularizer. It yields better recovery than both $\ell_0$ and $\ell_1$ regularizers on compressed sensing. However, how to efficiently optimize its learning problem is still challenging. The main difficulty is that both the $\ell_1$ and $\ell_2$ norms in $\ell_{1-2}$ are not differentiable, and existing optimization algorithms cannot be applied. In this paper, we show that a closed-form solution can be derived for the proximal step associated with this regularizer. We further extend the result for low-rank matrix learning and the total variation model. Experiments on both synthetic and real data sets show that the resultant accelerated proximal gradient algorithm is more efficient than other noncovex optimization algorithms.

cs.LG↗

arXiv2017

Efficient Inexact Proximal Gradient Algorithm for Nonconvex Problems

The proximal gradient algorithm has been popularly used for convex optimization. Recently, it has also been extended for nonconvex problems, and the current state-of-the-art is the nonmonotone accelerated proximal gradient algorithm. However, it typically requires two exact proximal steps in each iteration, and can be inefficient when the proximal step is expensive. In this paper, we propose an efficient proximal gradient algorithm that requires only one inexact (and thus less expensive) proximal step in each iteration. Convergence to a critical point %of the nonconvex problem is still guaranteed and has a $O(1/k)$ convergence rate, which is the best rate for nonconvex problems with first-order methods. Experiments on a number of problems demonstrate that the proposed algorithm has comparable performance as the state-of-the-art, but is much faster.

math.OC↗

arXiv2017

Multi-Label Learning with Global and Local Label Correlation

It is well-known that exploiting label correlations is important to multi-label learning. Existing approaches either assume that the label correlations are global and shared by all instances; or that the label correlations are local and shared only by a data subset. In fact, in the real-world applications, both cases may occur that some label correlations are globally applicable and some are shared only in a local group of instances. Moreover, it is also a usual case that only partial labels are observed, which makes the exploitation of the label correlations much more difficult. That is, it is hard to estimate the label correlations when many labels are absent. In this paper, we propose a new multi-label approach GLOCAL dealing with both the full-label and the missing-label cases, exploiting global and local label correlations simultaneously, through learning a latent label representation and optimizing label manifolds. The extensive experimental studies validate the effectiveness of our approach on both full-label and missing-label data.

cs.LG↗

arXiv2016

Stochastic Variance-Reduced ADMM

The alternating direction method of multipliers (ADMM) is a powerful optimization solver in machine learning. Recently, stochastic ADMM has been integrated with variance reduction methods for stochastic gradient, leading to SAG-ADMM and SDCA-ADMM that have fast convergence rates and low iteration complexities. However, their space requirements can still be high. In this paper, we propose an integration of ADMM with the method of stochastic variance reduced gradient (SVRG). Unlike another recent integration attempt called SCAS-ADMM, the proposed algorithm retains the fast convergence benefits of SAG-ADMM and SDCA-ADMM, but is more advantageous in that its storage requirement is very low, even independent of the sample size $n$. We also extend the proposed method for nonconvex problems, and obtain a convergence rate of $O(1/T)$. Experimental results demonstrate that it is as fast as SAG-ADMM and SDCA-ADMM, much faster than SCAS-ADMM, and can be used on much bigger data sets.

cs.LG↗

arXiv2016

Learning of Generalized Low-Rank Models: A Greedy Approach

Learning of low-rank matrices is fundamental to many machine learning applications. A state-of-the-art algorithm is the rank-one matrix pursuit (R1MP). However, it can only be used in matrix completion problems with the square loss. In this paper, we develop a more flexible greedy algorithm for generalized low-rank models whose optimization objective can be smooth or nonsmooth, general convex or strongly convex. The proposed algorithm has low per-iteration time complexity and fast convergence rate. Experimental results show that it is much faster than the state-of-the-art, with comparable or even better prediction performance.

cs.LG↗

arXiv2016

Fast Nonsmooth Regularized Risk Minimization with Continuation

In regularized risk minimization, the associated optimization problem becomes particularly difficult when both the loss and regularizer are nonsmooth. Existing approaches either have slow or unclear convergence properties, are restricted to limited problem subclasses, or require careful setting of a smoothing parameter. In this paper, we propose a continuation algorithm that is applicable to a large class of nonsmooth regularized risk minimization problems, can be flexibly used with a number of existing solvers for the underlying smoothed subproblem, and with convergence results on the whole algorithm rather than just one of its subproblems. In particular, when accelerated solvers are used, the proposed algorithm achieves the fastest known rates of $O(1/T^2)$ on strongly convex problems, and $O(1/T)$ on general convex problems. Experiments on nonsmooth classification and regression tasks demonstrate that the proposed algorithm outperforms the state-of-the-art.

cs.LG↗

arXiv2015

Asynchronous Distributed Semi-Stochastic Gradient Optimization

With the recent proliferation of large-scale learning problems,there have been a lot of interest on distributed machine learning algorithms, particularly those that are based on stochastic gradient descent (SGD) and its variants. However, existing algorithms either suffer from slow convergence due to the inherent variance of stochastic gradients, or have a fast linear convergence rate but at the expense of poorer solution quality. In this paper, we combine their merits by proposing a fast distributed asynchronous SGD-based algorithm with variance reduction. A constant learning rate can be used, and it is also guaranteed to converge linearly to the optimal solution. Experiments on the Google Cloud Computing Platform demonstrate that the proposed algorithm outperforms state-of-the-art distributed asynchronous algorithms in terms of both wall clock time and solution quality.

cs.LG↗

arXiv2015

Fast Low-Rank Matrix Learning with Nonconvex Regularization

Low-rank modeling has a lot of important applications in machine learning, computer vision and social network analysis. While the matrix rank is often approximated by the convex nuclear norm, the use of nonconvex low-rank regularizers has demonstrated better recovery performance. However, the resultant optimization problem is much more challenging. A very recent state-of-the-art is based on the proximal gradient algorithm. However, it requires an expensive full SVD in each proximal step. In this paper, we show that for many commonly-used nonconvex low-rank regularizers, a cutoff can be derived to automatically threshold the singular values obtained from the proximal operator. This allows the use of power method to approximate the SVD efficiently. Besides, the proximal operator can be reduced to that of a much smaller matrix projected onto this leading subspace. Convergence, with a rate of O(1/T) where T is the number of iterations, can be guaranteed. Extensive experiments are performed on matrix completion and robust principal component analysis. The proposed method achieves significant speedup over the state-of-the-art. Moreover, the matrix solution obtained is more accurate and has a lower rank than that of the traditional nuclear norm regularizer.

math.NA↗

arXiv2013

Convex and Scalable Weakly Labeled SVMs

In this paper, we study the problem of learning from weakly labeled data, where labels of the training examples are incomplete. This includes, for example, (i) semi-supervised learning where labels are partially known; (ii) multi-instance learning where labels are implicitly known; and (iii) clustering where labels are completely unknown. Unlike supervised learning, learning with weak labels involves a difficult Mixed-Integer Programming (MIP) problem. Therefore, it can suffer from poor scalability and may also get stuck in local minimum. In this paper, we focus on SVMs and propose the WellSVM via a novel label generation strategy. This leads to a convex relaxation of the original MIP, which is at least as tight as existing convex Semi-Definite Programming (SDP) relaxations. Moreover, the WellSVM can be solved via a sequence of SVM subproblems that are much more scalable than previous convex SDP relaxations. Experiments on three weakly labeled learning tasks, namely, (i) semi-supervised learning; (ii) multi-instance learning for locating regions of interest in content-based information retrieval; and (iii) clustering, clearly demonstrate improved performance, and WellSVM is also readily applicable on large data sets.

cs.LG↗

arXiv2013

Fast Stochastic Alternating Direction Method of Multipliers

In this paper, we propose a new stochastic alternating direction method of multipliers (ADMM) algorithm, which incrementally approximates the full gradient in the linearized ADMM formulation. Besides having a low per-iteration complexity as existing stochastic ADMM algorithms, the proposed algorithm improves the convergence rate on convex problems from $O(\frac 1 {\sqrt{T}})$ to $O(\frac 1 T)$, where $T$ is the number of iterations. This matches the convergence rate of the batch ADMM algorithm, but without the need to visit all the samples in each iteration. Experiments on the graph-guided fused lasso demonstrate that the new algorithm is significantly faster than state-of-the-art stochastic and batch ADMM algorithms.

cs.LG↗