Nanjing University of Science and Technology
Multiset Feature Learning for Highly Imbalanced Data Classification
Wu, Fei (Wuhan University and Nanjing University of Posts and Telecommunications) | Jing, Xiao-Yuan (Wuhan University and Nanjing University of Posts and Telecommunications) | Shan, Shiguang (Chinese Academy of Sciences (CAS)) | Zuo, Wangmeng (Harbin Institute of Technology) | Yang, Jing-Yu (Nanjing University of Science and Technology)
With the expansion of data, increasing imbalanced data has emerged. When the imbalance ratio of data is high, most existing imbalanced learning methods decline in classification performance. To address this problem, a few highly imbalanced learning methods have been presented. However, most of them are still sensitive to the high imbalance ratio. This work aims to provide an effective solution for the highly imbalanced data classification problem. We conduct highly imbalanced learning from the perspective of feature learning. We partition the majority class into multiple blocks with each being balanced to the minority class and combine each block with the minority class to construct a balanced sample set. Multiset feature learning (MFL) is performed on these sets to learn discriminant features. We thus propose an uncorrelated cost-sensitive multiset learning (UCML) approach. UCML provides a multiple sets construction strategy, incorporates the cost-sensitive factor into MFL, and designs a weighted uncorrelated constraint to remove the correlation among multiset features. Experiments on five highly imbalanced datasets indicate that: UCML outperforms state-of-the-art imbalanced learning methods.
Compressed K-Means for Large-Scale Clustering
Shen, Xiaobo (Nanjing University of Science and Technology) | Liu, Weiwei (University of Technology Sydney) | Tsang, Ivor (University of Technology Sydney) | Shen, Fumin (University of Electronic Science and Technology of China) | Sun, Quan-Sen (Nanjing University of Science and Technology)
Large-scale clustering has been widely used in many applications, and has received much attention. Most existing clustering methods suffer from both expensive computation and memory costs when applied to large-scale datasets. In this paper, we propose a novel clustering method, dubbed compressed k-means (CKM), for fast large-scale clustering. Specifically, high-dimensional data are compressed into short binary codes, which are well suited for fast clustering. CKM enjoys two key benefits: 1) storage can be significantly reduced by representing data points as binary codes; 2) distance computation is very efficient using Hamming metric between binary codes. We propose to jointly learn binary codes and clusters within one framework. Extensive experimental results on four large-scale datasets, including two million-scale datasets demonstrate that CKM outperforms the state-of-the-art large-scale clustering methods in terms of both computation and memory cost, while achieving comparable clustering accuracy.
Deep Learning with S-Shaped Rectified Linear Activation Units
Jin, Xiaojie (National University of Singapore) | Xu, Chunyan (Nanjing University of Science and Technology) | Feng, Jiashi (National University of Singapore) | Wei, Yunchao (National University of Singapore) | Xiong, Junjun (Beijing Samsung Telecom) | Yan, Shuicheng (National University of Singapore)
Rectified linear activation units are important components for state-of-the-art deep convolutional networks. In this paper, we propose a novel S-shaped rectifiedlinear activation unit (SReLU) to learn both convexand non-convex functions, imitating the multiple function forms given by the two fundamental laws, namely the Webner-Fechner law and the Stevens law, in psychophysics and neural sciences. Specifically, SReLU consists of three piecewise linear functions, which are formulated by four learnable parameters. The SReLU is learned jointly with the training of the whole deep network through back propagation. During the training phase, to initialize SReLU in different layers, we propose a “freezing” method to degenerate SReLU into a predefined leaky rectified linear unit in the initial several training epochs and then adaptively learn the good initial values. SReLU can be universally used in the existing deep networks with negligible additional parameters and computation cost. Experiments with two popular CNN architectures, Network in Network and GoogLeNet on scale-various benchmarks including CIFAR10, CIFAR100, MNIST and ImageNet demonstrate that SReLU achieves remarkable improvement compared to other activation functions.
Sparse Deep Stacking Network for Image Classification
Li, Jun (Nanjing University of Science and Technology) | Chang, Heyou (Nanjing University of Science and Technology) | Yang, Jian (Nanjing University of Science and Technology)
Sparse coding can learn good robust representation to noise and model more higher-order representation for image classification. However, the inference algorithm is computationally expensive even though the supervised signals are used to learn compact and discriminative dictionaries in sparse coding techniques. Luckily, a simplified neural network module (SNNM) has been proposed to directly learn the discriminative dictionaries for avoiding the expensive inference. But the SNNM module ignores the sparse representations. Therefore, we propose a sparse SNNM module by adding the mixed-norm regularization (l1/l2 norm). The sparse SNNM modules are further stacked to build a sparse deep stacking network (S-DSN). In the experiments, we evaluate S-DSN with four databases, including Extended YaleB, AR, 15 scene and Caltech101. Experimental results show that our model outperforms related classification methods with only a linear classifier. It is worth noting that we reach 98.8% recognition accuracy on 15 scene.
Intra-View and Inter-View Supervised Correlation Analysis for Multi-View Feature Learning
Jing, Xiao-Yuan (Wuhan University) | Hu, Rui-Min (Wuhan University) | Zhu, Yang-Ping (Nanjing University of Posts and Telecommunications) | Wu, Shan-Shan (Nanjing University of Posts and Telecommunications) | Liang, Chao (Wuhan University) | Yang, Jing-Yu (Nanjing University of Science and Technology)
Multi-view feature learning is an attractive research topic with great practical success. Canonical correlation analysis (CCA) has become an important technique in multi-view learning, since it can fully utilize the inter-view correlation. In this paper, we mainly study the CCA based multi-view supervised feature learning technique where the labels of training samples are known. Several supervised CCA based multi-view methods have been presented, which focus on investigating the supervised correlation across different views. However, they take no account of the intra-view correlation between samples. Researchers have also introduced the discriminant analysis technique into multi-view feature learning, such as multi-view discriminant analysis (MvDA). But they ignore the canonical correlation within each view and between all views. In this paper, we propose a novel multi-view feature learning approach based on intra-view and inter-view supervised correlation analysis (I2SCA), which can explore the useful correlation information of samples within each view and between all views. The objective function of I2SCA is designed to simultaneously extract the discriminatingly correlated features from both inter-view and intra-view. It can obtain an analytical solution without iterative calculation. And we provide a kernelized extension of I2SCA to tackle the linearly inseparable problem in the original feature space. Four widely-used datasets are employed as test data. Experimental results demonstrate that our proposed approaches outperform several representative multi-view supervised feature learning methods.
Learning Low-Rank Representations with Classwise Block-Diagonal Structure for Robust Face Recognition
Li, Yong (Chinese Academy of Sciences) | Liu, Jing (Chinese Academy of Sciences) | Li, Zechao (Nanjing University of Science and Technology) | Zhang, Yangmuzi (University of Maryland, College Park) | Lu, Hanqing (Chinese Academy of Sciences) | Ma, Songde (Chinese Academy of Sciences)
Face recognition has been widely studied due to its importance in various applications. However, the case that both training images and testing images are corrupted is not well addressed. Motivated by the success of low-rank matrix recovery, we propose a novel semi-supervised low-rank matrix recovery algorithm for robust face recognition. The proposed method can learn robust discriminative representations for both training images and testing images simultaneously by exploiting the classwise block-diagonal structure. Specifically, low-rank matrix approximation can handle the possible contamination of data. Moreover, the classwise block-diagonal structure is exploited to promote discrimination of representations for robust recognition. The above issues are formulated into a unified objective function and we design an efficient optimization procedure based on augmented Lagrange multiplier method to solve it. Extensive experiments on three public databases are performed to validate the effectiveness of our approach. The strong identification capability of representations with block-diagonal structure is verified.
Instance-Based Domain Adaptation in NLP via In-Target-Domain Logistic Approximation
Xia, Rui (Nanjing University of Science and Technology) | Yu, Jianfei (Nanjing University of Science and Technology) | Xu, Feng (Nanjing University of Science and Technology) | Wang, Shumei (Nanjing University of Science and Technology)
In the field of NLP, most of the existing domain adaptation studies belong to the feature-based adaptation, while the research of instance-based adaptation is very scarce. In this work, we propose a new instance-based adaptation model, called in-target-domain logistic approximation (ILA). In ILA, we adapt the source-domain data to the target domain by a logistic approximation. The normalized in-target-domain probability is assigned as an instance weight to each of the source-domain training data. An instance-weighted classification model is trained finally for the cross-domain classification problem. Compared to the previous techniques, ILA conducts instance adaptation in a dimensionality-reduced linear feature space to ensure efficiency in high-dimensional NLP tasks. The instance weights in ILA are learnt by leveraging the criteria of both maximum likelihood and minimum statistical distance. The empirical results on two NLP tasks including text categorization and sentiment classification show that our ILA model beats the state-of-the-art instance adaptation methods significantly, in cross-domain classification accuracy, parameter stability and computational efficiency.
What Will Others Choose? How a Majority Vote Reward Scheme Can Improve Human Computation in a Spatial Location Identification Task
Rao, Huaming (Nanjing University of Science and Technology) | Huang, Shih-Wen (University of Illinois at Urbana-Champaign) | Fu, Wai-Tat (University of Illinois at Urbana-Champaign)
We created a spatial location identification task (SpLIT) in which workers recruited from Amazon Mechanical Turk were presented with a camera view of a location, and were asked to identify the location on a two-dimensional map. In cases where these cues were ambiguous or did not provide enough information to pinpoint the exact location, workers had to make a best guess. We tested the effects of two reward schemes. In the “ground truth” scheme, workers were rewarded if their answers were close enough to the correct locations. In the “majority vote” scheme, workers were told that they would be rewarded if their answers were similar to the majority of other workers. Results showed that the majority vote reward scheme led to consistently more accurate answers. Cluster analysis further showed that the majority vote reward scheme led to answers with higher reliability (a higher percentage of answers in the correct clusters) and precision (a smaller average distance to the cluster centers). Possible reasons for why the majority voting reward scheme was better were discussed.