# robust_multiview_learning_via_halfquadratic_minimization__d21dcb54.pdf

Robust Multi-view Learning via Half-quadratic Minimization

Yonghua Zhu1,2, Xiaofeng Zhu1, , Wei Zheng1

1 Guangxi Key Lab of Multi-source Information Mining & Security, Guangxi Normal University, Guilin, 541004, China 2 Guangxi University, Nanning, 530004, China yhzhu66@qq.com, seanzhuxf@gmail.com, zwgxnu@163.com

Although multi-view clustering is capable to use more information than single view clustering, existing multi-view clustering methods still have issues to be addressed, such as initialization sensitivity, the speciﬁcation of the number of clusters, and the inﬂuence of outliers. In this paper, we propose a robust multi-view clustering method to address these issues. Speciﬁcally, we ﬁrst propose a multi-view based sum-of-square error estimation to make the initialization easy and simple as well as use a sum-of-norm regularization to automatically learn the number of clusters according to data distribution. We further employ robust estimators constructed by the half-quadratic theory to avoid the inﬂuence of outliers for conducting robust estimations of both sum-of-square error and the number of clusters. Experimental results on both synthetic and real datasets demonstrate that our method outperforms the state-of-the-art methods.

1 Introduction

Compared to single-view clustering, multi-view clustering is capable to generate more robust clustering models as every view of multi-view data has particular information different from other views so that multi-view data provide more useful information for clustering [Chi and Lange, 2015; Zhang et al., 2017; Nie et al., 2017; Zhu et al., 2014a]. Recently, reasonably taking full use of the information (e.g., common information and particular information) among different views has been demonstrated to be a key strategy for multi-view clustering beyond single view clustering [Kriegel et al., 2009; Hu et al., 2017; Rodriguez and Laio, 2014]. By considering the way of using the information in multiple views, existing multi-view clustering methods can be categorized into three categories, such as concatenation approach, distributed approach, and centralized approach. Concatenation approach ﬁrstly concatenates the feature vectors of every view to form a set of long vectors and then conducts single view clustering on the resulting long-vector representations. Existing concatenation approach is easy to perform, but

Corresponding author: Xiaofeng Zhu (seanzhuxf@gmail.com).

difﬁcultly yields satisﬁed clustering performance. For example, high-dimensional representation of long vectors in concatenation approach easily leads to the issue of curse of dimensionality [Li et al., 2017; Zhu et al., 2016]. Distributed approach ﬁrstly generates clustering results from every view independently, and then proposes strategies to combine these results. Similar to the concatenation approach, distributed approach does not take full use of information of multi-view data yet and thus is unavailable to yield reasonable clustering performance [Sun, 2013]. Centralized approach takes advantage of distinguishing information or common information among multi-view data to more easily achieve better performance, compared to the former two approaches [Yin et al., 2015]. However, centralized approach has well-known drawbacks such as initialization sensitivity [Shah and Koltun, 2017; Zhu et al., 2018]. Increasingly complex datasets and real-world requirements are creating new demands for multi-view clustering [Xu et al., 2016; Jiang et al., 2015]. Real-world data are often corrupted by outliers [Zhu et al., 2017b; Zhang et al., 2006], but most existing multi-view clustering methods were designed to equivalently regard every sample without addressing the issue of the inﬂuence of outliers. Another challenging task of multi-view clustering is the speciﬁcation of the number of clusters, which is required as an input. Usually, predeﬁned input needs prior knowledge. Otherwise, the obtained solution is prone to error, especially for conducting multi-view clustering on the datasets with complex distribution or structure [Frey and Dueck, 2007; Fang and Wang, 2012]. In this paper, we propose a novel Robust Multi-View Clustering (RMVC) method to conduct simple initialization sensitivity, taking sample importance into account to avoid the inﬂuence of outliers, and automatically specifying the number of the clusters. To do this, ﬁrstly, a sum-of-square error estimation is used to minimize the difference between original samples and their corresponding new representations, aim for resulting in easy initialization, i.e., initializing every new representation as its corresponding original feature to yield fast convergence. Moreover, we extend such a framework from single view data to multi-view data to further improve the clustering robustness by making the full use of the information in multi-view data. Meanwhile, we select robust estimators (based on the half-quadratic theory [Nikolova and Chan, 2007; He et al., 2014]) to push a constraint on every sample of

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

every view so that the outliers are removed out the model construction, unimportant samples are assigned small weights, and the important samples are assigned large weights. Secondly, a sum-of-norm regularization is used to control the tradeoff between the model ﬁt (i.e., the sum-of-square error estimation) and the number of clusters, so that the number of clusters can be automatically learnt according to the data distribution, and not need to be speciﬁed in advance. Compared to previous multi-view clustering methods, our proposed RMVC method has the following advantages.

Its initialization is easy as well as ﬁxed, i.e., initializing every sample with its original feature to achieve very simple and efﬁcient initialization. It is noteworthy that existing multi-view clustering methods (e.g., [Xu et al., 2015; Nie et al., 2017; Jiang et al., 2015; Cao et al., 2015]) are very sensitive to the initialization, and the initialization is complex such as a small set of random centroid is ﬁrstly initialized and then iteratively reﬁned [Chi and Lange, 2015; Zhu et al., 2014b]. Furthermore, different initializations lead to different clustering results. As a result, heuristic initialization methods should be used. Otherwise, clustering results may lead to a local minima and or even drastically suboptimal results.

It does not require the input of the number of clusters, i.e., automatically learning the number of clusters according to data distribution by a sum-of-norm regularization. By contrast, most of existing multi-view methods need to specify the number of clusters, which is time consuming as well as needs prior knowledge. Although some methods (e.g., afﬁnity propagation [Frey and Dueck, 2007; Zhu et al., 2017a]) do not need to assign the number of clusters by users, these methods need to face the initialization issue [Shah and Koltun, 2017; Hocking et al., 2011].

It is capable of simultaneously separating outliers from important samples and learning the number of clusters. Most of existing multi-view methods ignored these two constraints and only a few of previous clustering methods considered one of them. Moreover, our method is a general methodology that can be applied to other clustering problems such as single view clustering and semisupervised clustering.

We summarize all notations used in this paper in table 1.

2.1 Robust Clustering

K-means clustering partitions all the samples into K clusters so that every sample belongs to a speciﬁc cluster whose centroid is closest to this sample. Speciﬁcally, the objective function of k-means clustering [Ding et al., 2005] can be formulated as:

min G,F X GF 2 F

s.t., Gi,k {0, 1}, K P

k=1 Gi,k = 1, i = 1, ..., n (1)

X a matrix x a vector of X xi,j the element in the i-th row and the j-th column of X xi, the i-th row of X x ,j the j-th column of X

||X||F the Frobenius norm of X, i.e., ||X||F = q P

||X||2,1 the ℓ2,1-norm of X, i.e., ||X||2,1 = P

j x2 ij XT the transpose of X tr(X) the trace of X X 1 the inverse of X

Table 1: The detail of the notations used in this paper.

where G Rn k is a binary matrix to indicate the samples belonging to a speciﬁc cluster out of K clusters, and F Rk d is the centroid matrix, and F is a Frobenius norm. Firstly, k-means clustering is sensitive to initialization since it is a non-convex optimization problem, whose optimization easily leads to a local minima or even sometimes drastically suboptimal. Moreover, heuristic initialization methods should be used for k-means clustering since different initializations may lead to considerably different clustering results [Hocking et al., 2011]. Secondly, k-means clustering needs to set the number of clusters by users. It is usually time consuming for selecting an appropriate value for the number of clusters and there is a demand of prior knowledge. Thirdly, most existing clustering methods ignore the inﬂuence of outliers so that constructed clustering models deviate to the real ones. Intuitively, closed samples in some similarity/distance measures should be assigned to the same cluster, and vice versa. As a result, a centroid can be obtained for each cluster. In real applications, we have no idea on the number of clusters, so we denote ui, (the i-th row of U Rn d, i = 1, ..., n) as the new representation of the sample xi, as well as the centroid containing the sample xi, . We expect to ﬁrstly initialize ui, as xi, , and then recursively join the closed samples together to form a cluster/group until all the samples are jointed their closest clusters. This results in a sum-of-square error as follows:

i=1 xi, ui, 2 2 (2)

where 2 is an ℓ2-norm. The minimization optimization of Eq. (2) easily results in trivial solutions, i.e., xi, = ui, (i = 1, ..., n). To address this issue, an appropriate constraint (e.g., an ℓp-norm regularization) can be added to have the following formulation:

i=1 xi, ui, 2 2 + λ X

(p,q) ε wp,q up, uq, p (3)

where p is an ℓp-norm regularization, (p, q) ε indicate that the p-th sample xp, is one of the nearest neighbors of the q-th sample xq, , i.e., their distance is less then ε, and λ is a tuning parameter balancing two terms of Eq. (3).

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

In Eq. (3), the ﬁrst term (i.e., the sum-of-square error estimation) learns the new representation U of X so that recursively updating U which is gradually close to X as well as satisﬁes the constraint in the second term, i.e., the sumof-norm regularization. The second term requires that U preserves the similarity of X and the new representations of samples from the same cluster are very approximate. Speciﬁcally, the similarity of X is preserved by an adjacency matrix W = [wp,q]p,q=1,...,n Rn n, which can be constructed by the k Nearest Neighbor (k NN) method, or the mutual k NN method [Brito et al., 1997], or the sparse representation method [Elhamifar and Vidal, 2013]. On the other hand, if the p-th sample and the q-th sample come the same clustering, i.e., wp,q ε, then their centroid should be same, i.e., up, = uq, . By inducing the ℓp-norm regularization and tuning the parameter λ, some same-cluster-pairs (i.e., the pair i j if ui, = uj, , i, j = 1, ...n) are generated. Moreover, the larger the value of λ, the more the number of samecluster-pairs is. After the optimization process, all the samples are gathered into clusters, and thus the number of the cluster is decided. Eq. (3) solves the issue of initialization sensitivity, but does not solve the issue of the speciﬁcation of the number of clusters well as the ℓp-norm regularization has still limitations on robustly dealing with the outlier edges of W, high-order outliers for short. That is, the outlier edges will still appear instead of diminishing during the process of optimizing U. Moreover, Eq. (3) does not touch the issue of outlier samples (low-order outliers for short) in its ﬁrst term. In this paper, we have two considerations. Firstly, the outlier edges in W will be suppressed via assigning a small or even zero weight on the difference estimation of up, uq, . Secondly, the outliers in X will also be assigned small weights. To do this, we employ robust estimators constructed by the half-quadratic theory to have:

i=1 ρ1( xi, ui, 2) + λ X

(p,q) ε wp,qρ2( up, uq, 2)

(4) where ρ1( ) and ρ2( ) are robust estimators constructed by the half-quadratic theory [Nikolova and Ng, 2005; Charbonnier et al., 1997; Black and Rangarajan, 1996]. Robust estimators were designed to produce robust estimation that are not unduly affected by outliers and output good performance when there are small departures from parametric distributions [Nikolova and Ng, 2005]. The popular robust estimators include ℓ1 ℓ2 estimator, Cauchy estimator, Geman-Mc Clure estimator, and so on.

2.2 Robust Multi-view Clustering

We denote Xv Rn dv of multi-view data as the feature matrix of the v-th view data, where n and dv, respectively, are the number of samples and the features of the v-th view data. We also denote xv i,j, xv i, and xv ,j, respectively, as the element of the i-th row and the j-th column, the vector of the i-th row, and the vector of the j-th column, of the v-th feature matrix. It is noteworthy that the dimensions of different views are usually different. In this paper, we ﬁrst map the feature

matrix Xv of each view into a kernel space ˆXv Rn n via a linear kernel. Thus, we extend Eq. (4) into the following Robust Multi-View Clustering (RMVC):

i=1 ρ1( ˆxv i, ˆui, 2)

(p,q) ε ˆwv p,qρ2( ˆup, ˆuq, 2) (5)

In Eq. (5), we expect to learn a common representation ˆU Rn n from all m views and save the similarity of every view in ˆ Wv (v = 1, ..., n) which are constructed by the mutual k NN method in this paper. After optimizing Eq. (5), we use the obtained ˆU (i.e., the common representation of multi-view data) to conduct multi-view clustering. For simplicity, in this paper, we use the ℓ1 ℓ2 estimator to replace both ρ1( ) and ρ2( ), i.e., ρ1(z) = p

γ2 + z2 (6)

where z is a vector or matrix variable, γ1 and γ2 are used to control the number of samples involving into the process of optimization in each iteration. Hence, we obtain the ﬁnal objective function as follows:

γv 1 + ˆxv i, ˆui, 2 2

(p,q) ε ˆwv p,q p

γ2 + ˆup, ˆuq, 2 2 (7)

where γv 1 is the tuning parameter of the v-th view data ˆXv. For efﬁcient optimization and the control of the number of samples involving into the process of optimization in each iteration, the half-quadratic theory usually transfers the robust estimator into its equivalent formulation via adding an extra variable c, i.e., min z ρ(z) min c,z cz2 + ψ(c), where cz2 is

a quadratic function, ψ( ) is the dual function of the robust estimator [He et al., 2014]. Usually, different robust estimators have different dual functions and every dual function has its minimization function which decides the weight vector c. After such a transformation, each element ci in the weight vector c is the weight of the corresponding zi. Speciﬁcally, if zi is an outlier, then the value of ci is small or even 0, and vice versa. Hence, the inﬂuence of outliers is suppressed or even diminished while tuning the parameters of robust estimators, such as γv 1 and γ2 in Eq. (8). Moreover, we do not need to know the explicit function of the dual function as c is decided by the minimization function of the dual function. Based on above analysis, we transfer our objective function in Eq. (7) to the following problem:

min ˆ U,αv,β

i=1 αv i ˆxv i, ˆui, 2 2 + ψ1(αv))

(p,q) ε ˆwv p,qβp,q ˆup, ˆuq, 2 2 + ψ2(β) (8)

where αv Rn 1 and β Rn n. The vector αv is used to control outlier samples in Xv and β is used to control the outlier edges in ˆ Wv.

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

Eq. (8) is not convex for all the variables (i.e., U, αv (v = 1, ..., m), β) but is convex for each of them while ﬁxing the others. In this paper, we employ the alternative optimization strategy to optimize Eq. (8), i.e., (i) optimizing αv and β while ﬁxing U and (ii) optimizing U while ﬁxing αv and β. We literately repeat the step (i) and (ii) until the objective value achieves stable or satisﬁes the predeﬁned demand of stop criteria.

2.3 Convergence To prove the convergence of optimization algorithm (i.e., the alternative strategy), we ﬁrst denote ˆU(t+1) as the value of (t +1)-th iteration of ˆU. Thus Eq. (8) is changed into:

J(αv(t+1), β(t+1), ˆU(t+1)) = m P

i=1 αv i (t+1) ˆxv i, ˆu(t+1) i, 2 2 + ψ1(αv(t+1)))+

(p,q) ε ˆwv p,qβ(t+1) p,q ˆu(t+1) p, ˆu(t+1) q, 2 2 + ψ2(β(t+1))

(9) As the variables αv and β are decided by the half-quadratic theory, whose convergence has been proved in [Nikolova and Ng, 2005]. Thus we have

J(αv(t+1), β(t+1), ˆU(t+1)) J(αv(t), β(t), ˆU(t+1)) (10) The variable ˆU has a closed form solution, so

J(αv(t), β(t), ˆU(t+1)) J(αv(t), β(t), ˆU(t)) (11) By plugging Eq. (11) into Eq. (10), we have:

J(αv(t+1), β(t+1), ˆU(t+1)) J(αv(t), β(t), ˆU(t)) (12) Hence, optimization algorithm converges to a local minima.

3 Experiments We evaluated our proposed method by comparing with three state-of-the-art multi-view clustering and a single view clustering method on one synthetic multi-view dataset and four real multi-view datasets in term of four clustering metrics, such as ACCuracy of clustering (ACC), Normalized Mutual Information (NMI), Purity, and Adjusted Rand Index (ARI).

3.1 Experimental Settings In our experiments, we evaluated our method by comparing with ﬁve comparison methods, which details were summarized as follows: We conducted k-means clustering on every view of multi-view data and denoted the worst result and best result, respectively, as Worst and Best. As a concatenation approach, Concatenation k-means clustering (Con KM) concatenates the features cross all the views to form a long vector and then conducts kmeans clustering on a single dataset. As a distributed approach, Weighted view Collaborative Fuzzy C-means (Wv Co FCM) [Jiang et al., 2015] ﬁrst conducts clustering on each view based on the fuzzy k-means clustering and then combines all the resulting result to output the ﬁnal clustering result. Wv Co FCM takes the importance of each view into account.

Data Sets #(Views) #(Samples) #(Classes) #(Type) Cornell 4 195 5 Web Texas 4 187 5 Web Washingdon 4 230 5 Web Wisconsin 4 265 5 Web

Table 2: The details of public multi-view datasets.

As a centralized approach, Multi-View Self-Paced Learning for Clustering (MSPLC) [Xu et al., 2015] uses self-paced learning to simultaneously consider both the sample-diversity and the view-diversity. In our experiments, we repeated k-means clustering 20 times and reported their average value for single view clustering and Con KM, and set the value of k as [10, 15, ..., 40] in k-nearest neighbor for the methods such as our proposed method. We also set the number of clusters as the number of real classes for all comparison methods, while our method automatically generates the number of clusters. The ranges of parameters of every method were set by strictly following the corresponding literature. We set the range of parameter λ in our method as λ [10 3, 10 2, ..., 103] and set the stop criteria of the iterative optimization as obj(t+1) obj(t) 2 2 obj(t) 10 6, where obj(t) stands for the objective value in the t-th iteration.

3.2 Synthetic Datasets The synthetic dataset consisted of ﬁve letters (such as I , J , C , A , and I ), which were represented by ﬁve views. The samples in each letter formed a cluster and were visualized with a color different from others. The ﬁrst letter and the last letter were far away other letters in ﬁrst two views, the former two letters and the last two letters were far away from other letters in other two views, and the middle letter was far away from other letters in 5th view. Figure 1 visualized the clustering results of all the methods, where all the clustering methods did not correctly output exact clustering results since the shapes of clusters are different and irregular. However, our method achieved the best clustering performance, compared to all the comparison methods. Actually, it is difﬁcult for the comparison methods to have effective initializations on complex multi-view datasets. For example, random sampling for the initialization was unavailable to obtain initial centroid cross all the clusters, so the ﬁnal clustering results easily achieved local minima or even sub-optimality. By contrast, our proposed method exactly selected all centroid since the initialization with original samples preserved the structures of the data better than random sampling methods. Thus, our method could easily achieve good clustering performance with simple initialization. In a word, our method outperformed all the comparison methods for multi-view clustering on the synthetic dataset, which was with predeﬁned number of clusters. The experimental results veriﬁed the advantages of our method in term of initialization sensitivity via the synthetic dataset.

3.3 Real Datasets We used four real multi-view datasets to evaluate the practical clustering performance of our proposed method. The details of the used datasets are reported in Table 2 and the clustering

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

(d) Wv Co FCM

(f) Proposed

Figure 1: Cluster results of the synthetic dataset.

results of all the methods on these datasets were reported in Table 3. Obviously, our method achieved the best clustering performance, followed by Wv Co FCM, MSPLC and Con KM. For example, the accuracy of our method improved on average by 8.71% and 12.88%, respectively, compared to the best comparison method (i.e., Wv Co FCM) and the worst comparison method (i.e., Con KM) on all four datasets. The possible reason is that our method simultaneously takes three constraints into account, i.e., initialization sensitivity, the inﬂuences of outliers, and the speciﬁcation of the number of clusters. By contrast, all the comparison methods only consider a part of these constraints, so that were unavailable to output robust clustering models. In particular, both our proposed method and MSPLC considered the inﬂuence of outliers, but MSPLC does not touch other constraints and only took the inﬂuence of the low-order outliers into account.

3.4 Parameters Tuning

In this section, we investigated the variations of our method with different parameters setting. Firstly, we conducted experiments on our method with different parameters (k and λ) and listed the ACC results in Figure 2. The results showed that our method is sensitive a little to parameters setting. For example, our method easily achieved the best clustering performance while setting k [10, 20] and λ [101, 103] on dataset Wisconsin. This indicates that the parameters ranges of our method can be easily adjusted to achieve reasonable clustering performance. Secondly, we experimentally demonstrated the convergence of our proposed optimization algorithm to solve our objective function Eq. (7) via listing the variations of objective function values in Eq. (7) with different numbers of iterations in Figure 3. As a result, our proposed optimization algorithm fast converged within 20 iterations.

Thirdly, we conducted experiments to show the feasibility of automatically generating the number of clusters. To do this, we generated different binary matrices via tuning the parameters in our method so that our method outputted different numbers of clusters. We reported the results in Figure 4, where yellow color is the results while the number of clusters is the real classes. From Figure 4, the datasets partitioned into different numbers of clusters may yield different clustering results. Moreover, the yellow results may be worse than other results. This indicates that the real class may not be the best number of clusters. This is true in real applications because one cluster may be partitioned into multiple sub-clusters and some clusters may also be categorized into one cluster [Zhu and Martinez, 2006]. Moreover, the results in Figure 4 veriﬁed the advantages of our proposed method in terms of the speciﬁcation of the number of clusters, i.e., ﬂexibly outputting the number of clusters according to data distribution and not needing to preset the number of clusters for clustering.

4 Conclusion In this paper, we have proposed a novel robust multi-view clustering method to deal with the issues of existing multiview clustering, such as initialization sensitivity, the inﬂuence of outliers, and the speciﬁcation of the number of clusters. Experimental results on both synthetic and real datasets veriﬁed that these constraints affect clustering results as our method considering all of them outperformed the comparison methods only considering a part of these constraints.

Acknowledgements This work was supported in part by the Nation Natural Science Foundation of China (Grants No: 61573270 and 61672177), the China Key Research Program (Grant No: 2016YFB1000905), the Guangxi Natural Science Foundation

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

Datasets Cornell Texas Washingdon Wisconsin ACC NMI Purity ARI ACC NMI Purity ARI ACC NMI Purity ARI ACC NMI Purity ARI Worst 0.3641 0.0578 0.4308 0.0394 0.4439 0.2086 0.6043 0.0768 0.4478 0.0278 0.4652 0.0084 0.4113 0.0328 0.4679 0.0230 Best 0.4256 0.0274 0.4359 0.0081 0.5561 0.1506 0.5829 0.1673 0.4739 0.0259 0.4826 0.0076 0.5208 0.1614 0.5396 0.1808 Con KM 0.4359 0.2709 0.5744 0.1336 0.4492 0.2048 0.6086 0.1297 0.4565 0.0148 0.5696 0.0057 0.4906 0.0496 0.4981 0.0181 WVCo FCM 0.4513 0.3032 0.5897 0.1499 0.5027 0.3382 0.6952 0.2287 0.4913 0.2925 0.6565 0.2016 0.5509 0.1827 0.5547 0.1889 MSPLC 0.3995 0.2288 0.4462 0.0167 0.4385 0.2568 0.5615 0.0271 0.4774 0.3216 0.6470 0.2508 0.5102 0.3445 0.5509 0.0779 Proposed 0.4945 0.3125 0.6006 0.1482 0.6087 0.3215 0.7067 0.2378 0.6478 0.3642 0.7054 0.3379 0.5964 0.3538 0.6304 0.1791

Table 3: Clustering results on real multi-view data sets.

0 7 -1 8 -2 -3

(a) Cornell

0 7 -1 8 -2 -3

0 7 -1 8 -2 -3

(c) Washingdon

0 2 -1 3 -2 -3

(d) Wisconsin

Figure 2: The variation of different parameter combination.

0 10 20 Iterations

(a) Cornell

0 10 20 Iterations

0 10 20 Iterations

(c) Washingdon

0 10 20 Iterations

(d) Wisconsin

Figure 3: The Objective Function Value (OFV) of our method with different iterations.

3 4 5 6 7 8 9 Number of classes

(a) Cornell

2 3 4 5 6 7 8 Number of classes

4 5 6 7 8 9 10 Number of classes

(c) Washingdon

2 3 4 5 6 7 8 Number of classes

(d) Wisconsin

Figure 4: ACC of our method with different numbers of clusters, where yellow-color represented the clustering results with the real classes.

(Grant No: 2015GXNSFCB139011), the Guangxi High Institutions Program of Introducing 100 High-Level Overseas Talents, the Guangxi Collaborative Innovation Center of Multi Source Information Integration and Intelligent Processing, and the Research Fund of Guangxi Key Lab of Multi-source Information Mining & Security (18-A-01-01).

[Black and Rangarajan, 1996] Michael J. Black and Anand Rangarajan. On the uniﬁcation of line processes, outlier rejection, and robust statistics with applications in early vision. International Journal of Computer Vision, 19(1):57 91, 1996.

[Brito et al., 1997] MR Brito, EL Chavez, AJ Quiroz, and JE Yukich. Connectivity of the mutual k-nearest-neighbor graph in clustering and outlier detection. Statistics & Probability Letters, 35(1):33 42, 1997. [Cao et al., 2015] Xiaochun Cao, Changqing Zhang, Huazhu Fu, Si Liu, and Hua Zhang. Diversity-induced multi-view subspace clustering. In CVPR, pages 586 594, 2015. [Charbonnier et al., 1997] Pierre Charbonnier, Laure Blanc F eraud, Gilles Aubert, and Michel Barlaud. Deterministic edge-preserving regularization in computed imaging. IEEE Trans. Image Processing, 6(2):298 311, 1997. [Chi and Lange, 2015] Eric C Chi and Kenneth Lange. Splitting methods for convex clustering. Journal of Computational and Graphical Statistics, 24(4):994 1013, 2015.

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)

[Ding et al., 2005] Chris Ding, Xiaofeng He, and Horst D Simon. On the equivalence of nonnegative matrix factorization and spectral clustering. In SDMs, pages 606 610, 2005. [Elhamifar and Vidal, 2013] Ehsan Elhamifar and Rene Vidal. Sparse subspace clustering: Algorithm, theory, and applications. IEEE transactions on pattern analysis and machine intelligence, 35(11):2765 2781, 2013. [Fang and Wang, 2012] Yixin Fang and Junhui Wang. Selection of the number of clusters via the bootstrap method. Computational Statistics & Data Analysis, 56(3):468 477, 2012. [Frey and Dueck, 2007] Brendan J Frey and Delbert Dueck. Clustering by passing messages between data points. science, 315(5814):972 976, 2007. [He et al., 2014] Ran He, Wei-Shi Zheng, Tieniu Tan, and Zhenan Sun. Half-quadratic-based iterative minimization for robust sparse representation. IEEE transactions on pattern analysis and machine intelligence, 36(2):261 275, 2014. [Hocking et al., 2011] Toby Dylan Hocking, Armand Joulin, Francis Bach, and Jean-Philippe Vert. Clusterpath an algorithm for clustering using convex fusion penalties. In ICML, page 1, 2011. [Hu et al., 2017] Rongyao Hu, Xiaofeng Zhu, Debo Cheng, Wei He, Yan Yan, Jingkuan Song, and Shichao Zhang. Graph self-representation method for unsupervised feature selection. Neurocomputing, 220:130 137, 2017. [Jiang et al., 2015] Yizhang Jiang, Fu-Lai Chung, Shitong Wang, Zhaohong Deng, Jun Wang, and Pengjiang Qian. Collaborative fuzzy clustering from multiple weighted views. IEEE Trans. Cybernetics, 45(4):688 701, 2015. [Kriegel et al., 2009] Hans-Peter Kriegel, Peer Kr oger, and Arthur Zimek. Clustering high-dimensional data: A survey on subspace clustering, pattern-based clustering, and correlation clustering. TKDD, 3(1):1:1 1:58, 2009. [Li et al., 2017] Jundong Li, Kewei Cheng, Suhang Wang, Fred Morstatter, Robert P. Trevino, Jiliang Tang, and Huan Liu. Feature selection: A data perspective. ACM Comput. Surv., 50(6):94:1 94:45, 2017. [Nie et al., 2017] Feiping Nie, Guohao Cai, and Xuelong Li. Multi-view clustering and semi-supervised classiﬁcation with adaptive neighbours. In AAAI, pages 2408 2414, 2017. [Nikolova and Chan, 2007] Mila Nikolova and Raymond H. Chan. The equivalence of half-quadratic minimization and the gradient linearization iteration. IEEE Trans. Image Processing, 16(6):1623 1627, 2007. [Nikolova and Ng, 2005] Mila Nikolova and Michael K. Ng. Analysis of half-quadratic minimization methods for signal and image recovery. SIAM J. Scientiﬁc Computing, 27(3):937 966, 2005. [Rodriguez and Laio, 2014] Alex Rodriguez and Alessandro Laio. Clustering by fast search and ﬁnd of density peaks. Science, 344(6191):1492 1496, 2014.

[Shah and Koltun, 2017] Sohil Atul Shah and Vladlen Koltun. Robust continuous clustering. Proceedings of the National Academy of Sciences of the United States of America, 114(37):9814, 2017. [Sun, 2013] Shiliang Sun. A survey of multi-view machine learning. Neural Computing and Applications, 23(78):2031 2038, 2013. [Xu et al., 2015] Chang Xu, Dacheng Tao, and Chao Xu. Multi-view self-paced learning for clustering. In IJCAI, pages 3974 3980, 2015. [Xu et al., 2016] Jinglin Xu, Junwei Han, Kai Xiong, and Feiping Nie. Robust and sparse fuzzy k-means clustering. In IJCAI, pages 2224 2230, 2016. [Yin et al., 2015] Qiyue Yin, Shu Wu, Ran He, and Liang Wang. Multi-view clustering via pairwise sparse subspace representation. Neurocomputing, 156:12 21, 2015. [Zhang et al., 2006] Shichao Zhang, Yongsong Qin, Xiaofeng Zhu, Jilian Zhang, and Chengqi Zhang. Optimized parameters for missing data imputation. In PRICAI, pages 1010 1016, 2006. [Zhang et al., 2017] Shichao Zhang, Xuelong Li, Ming Zong, Xiaofeng Zhu, and Ruili Wang. Efﬁcient knn classiﬁcation with different numbers of nearest neighbors. IEEE transactions on neural networks and learning systems,DOI: 10.1109/TNNLS.2017.2673241, 2017. [Zhu and Martinez, 2006] Manli Zhu and Aleix M Martinez. Subclass discriminant analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(8):1274 1286, 2006. [Zhu et al., 2014a] Xiaofeng Zhu, Heung-Il Suk, and Dinggang Shen. Multi-modality canonical feature selection for alzheimer s disease diagnosis. In MICCAI, pages 162 169, 2014. [Zhu et al., 2014b] Yingying Zhu, Dong Huang, Fernando De La Torre, and Simon Lucey. Complex non-rigid motion 3d reconstruction by union of subspaces. In CVPR, pages 1542 1549, 2014. [Zhu et al., 2016] Yingying Zhu, Xiaofeng Zhu, Minjeong Kim, Dinggang Shen, and Guorong Wu. Early diagnosis of alzheimer s disease by joint feature selection and classiﬁcation on temporally structured support vector machine. In MICCAI, pages 264 272, 2016. [Zhu et al., 2017a] Pengfei Zhu, Wencheng Zhu, Qinghua Hu, Changqing Zhang, and Wangmeng Zuo. Subspace clustering guided unsupervised feature selection. Pattern Recognition, 66:364 374, 2017. [Zhu et al., 2017b] Yingying Zhu, Xiaofeng Zhu, Minjeong Kim, Daniel Kaufer, and Guorong Wu. A novel dynamic hyper-graph inference framework for computer assisted diagnosis of neuro-diseases. In IPMI, pages 158 169. Springer, 2017. [Zhu et al., 2018] Pengfei Zhu, Qian Xu, Qinghua Hu, Changqing Zhang, and Hong Zhao. Multi-label feature selection with missing labels. Pattern Recognition, 74:488 502, 2018.

Proceedings of the Twenty-Seventh International Joint Conference on Artiﬁcial Intelligence (IJCAI-18)