Domain adaptation from multiple sources: a domain-dependent regularization approach

Flowing on Riemannian manifold: domain adaptation by shifting covariance

Domain adaptation has shown promising results in computer vision applications. In this paper, we propose a new unsupervised domain adaptation method called domain adaptation by shifting covariance (DASC) for object recognition without requiring any labeled samples from the target domain. By characterizing samples from each domain as one covariance matrix, the source and target domain are represented into two distinct points residing on a Riemannian manifold. Along the geodesic constructed from the two points, we then interpolate some intermediate points (i.e., covariance matrices), which are used to bridge the two domains. By utilizing the principal components of each covariance matrix, samples from each domain are further projected into intermediate feature spaces, which finally leads to domain-invariant features after the concatenation of these features from intermediate points. In the multiple source domain adaptation task, we also need to effectively integrate different types of features between each pair of source and target domains. We additionally propose an SVM based method to simultaneously learn the optimal target classifier as well as the optimal weights for different source domains. Extensive experiments demonstrate the effectiveness of our method for both single source and multiple source domain adaptation tasks.

Generalized hidden-mapping ridge regression, knowledge-leveraged inductive transfer learning for neural networks, fuzzy systems and kernel methods

Inductive transfer learning has attracted increasing attention for the training of effective model in the target domain by leveraging the information in the source domain. However, most transfer learning methods are developed for a specific model, such as the commonly used support vector machine, which makes the methods applicable only to the adopted models. In this regard, the generalized hidden-mapping ridge regression (GHRR) method is introduced in order to train various types of classical intelligence models, including neural networks, fuzzy logical systems and kernel methods. Furthermore, the knowledge-leverage based transfer learning mechanism is integrated with GHRR to realize the inductive transfer learning method called transfer GHRR (TGHRR). Since the information from the induced knowledge is much clearer and more concise than that from the data in the source domain, it is more convenient to control and balance the similarity and difference of data distributions between the source and target domains. The proposed GHRR and TGHRR algorithms have been evaluated experimentally by performing regression and classification on synthetic and real world datasets. The results demonstrate that the performance of TGHRR is competitive with or even superior to existing state-of-the-art inductive transfer learning algorithms.

Learning With Augmented Features for Supervised and Semi-Supervised Heterogeneous Domain Adaptation

In this paper, we study the heterogeneous domain adaptation (HDA) problem, in which the data from the source domain and the target domain are represented by heterogeneous features with different dimensions. By introducing two different projection matrices, we first transform the data from two domains into a common subspace such that the similarity between samples across different domains can be measured. We then propose a new feature mapping function for each domain, which augments the transformed samples with their original features and zeros. Existing supervised learning methods (e.g., SVM and SVR) can be readily employed by incorporating our newly proposed augmented feature representations for supervised HDA. As a showcase, we propose a novel method called Heterogeneous Feature Augmentation (HFA) based on SVM. We show that the proposed formulation can be equivalently derived as a standard Multiple Kernel Learning (MKL) problem, which is convex and thus the global solution can be guaranteed. To additionally utilize the unlabeled data in the target domain, we further propose the semi-supervised HFA (SHFA) which can simultaneously learn the target classifier as well as infer the labels of unlabeled target samples. Comprehensive experiments on three different applications clearly demonstrate that our SHFA and HFA outperform the existing HDA methods.

Transfer ordinal label learning

Designing a classifier in the absence of labeled data is becoming a common encounter as the acquisition of informative labels is often difficult or expensive, particularly on new uncharted target domains. The feasibility of attaining a reliable classifier for the task of interest is embarked by some in transfer learning, where label information from relevant source domains is considered for complimenting the design process. The core challenge arising from such endeavors, however, is the induction of source sample selection bias, such that the trained classifier has the tendency of steering toward the distribution of the source domain. In addition, this bias is deemed to become more severe on data involving multiple classes. Considering this cue, our interest in this paper is to address such a challenge in the target domain, where ordinal labeled data are unavailable. In contrast to the previous works, we propose a transfer ordinal label learning paradigm to predict the ordinal labels of target unlabeled data by spanning the feasible solution space with ensemble of ordinal classifiers from the multiple relevant source domains. Specifically, the maximum margin criterion is considered here for the construction of the target classifier from an ensemble of source ordinal classifiers. Theoretical analysis and extensive empirical studies on real-world data sets are presented to study the benefits of the proposed method.

Hinging hyperplanes for time-series segmentation

Division of a time series into segments is a common technique for time-series processing, and is known as segmentation. Segmentation is traditionally done by linear interpolation in order to guarantee the continuity of the reconstructed time series. The interpolation-based segmentation methods may perform poorly for data with a level of noise because interpolation is noise sensitive. To handle the problem, this paper establishes an explicit expression for segmentation from a compact representation for piecewise linear functions using hinging hyperplanes. This expression enables the use of regression to obtain a continuous reconstructed signal and, as a consequence, application of advanced techniques in segmentation. In this paper, a least squares support vector machine with lasso using a hinging feature map is given and analyzed, based on which a segmentation algorithm and its online version are established. Numerical experiments conducted on synthetic and real-world datasets demonstrate the advantages of our methods compared to existing segmentation algorithms.

Knowledge-leverage-based TSK Fuzzy System modeling

Classical fuzzy system modeling methods consider only the current scene where the training data are assumed to be fully collectable. However, if the data available from the current scene are insufficient, the fuzzy systems trained by using the incomplete datasets will suffer from weak generalization capability for the prediction in the scene. In order to overcome this problem, a knowledge-leverage-based fuzzy system (KL-FS) is studied in this paper from the perspective of transfer learning. The KL-FS intends to not only make full use of the data from the current scene in the learning procedure, but also effectively leverage the existing knowledge from the reference scenes. Specifically, a knowledge-leverage-based Takagi-Sugeno-Kang-type Fuzzy System (KL-TSK-FS) is proposed by integrating the corresponding knowledge-leverage mechanism. The new fuzzy system modeling technique is evaluated through experiments on synthetic and real-world datasets. The results demonstrate that KL-TSK-FS has better performance and adaptability than the traditional fuzzy modeling methods in scenes with insufficient data.

Stopped object detection by learning foreground model in videos

The automatic detection of objects that are abandoned or removed in a video scene is an interesting area of computer vision, with key applications in video surveillance. Forgotten or stolen luggage in train and airport stations and irregularly parked vehicles are examples that concern significant issues, such as the fight against terrorism and crime, and public safety. Both issues involve the basic task of detecting static regions in the scene. We address this problem by introducing a model-based framework to segment static foreground objects against moving foreground objects in single view sequences taken from stationary cameras. An image sequence model, obtained by learning in a self-organizing neural network image sequence variations, seen as trajectories of pixels in time, is adopted within the model-based framework. Experimental results on real video sequences and comparisons with existing approaches show the accuracy of the proposed stopped object detection approach.

Developing a local least-squares support vector machines-based neuro-fuzzy model for nonlinear and chaotic time series prediction

Local modeling approaches, owing to their ability to model different operating regimes of nonlinear systems and processes by independent local models, seem appealing for modeling, identification, and prediction applications. In this paper, we propose a local neuro-fuzzy (LNF) approach based on the least-squares support vector machines (LSSVMs). The proposed LNF approach employs LSSVMs, which are powerful in modeling and predicting time series, as local models and uses hierarchical binary tree (HBT) learning algorithm for fast and efficient estimation of its parameters. The HBT algorithm heuristically partitions the input space into smaller subdomains by axis-orthogonal splits. In each partitioning, the validity functions automatically form a unity partition and therefore normalization side effects, e.g., reactivation, are prevented. Integration of LSSVMs into the LNF network as local models, along with the HBT learning algorithm, yield a high-performance approach for modeling and prediction of complex nonlinear time series. The proposed approach is applied to modeling and predictions of different nonlinear and chaotic real-world and hand-designed systems and time series. Analysis of the prediction results and comparisons with recent and old studies demonstrate the promising performance of the proposed LNF approach with the HBT learning algorithm for modeling and prediction of nonlinear and chaotic systems and time series.

Robust support vector regression for uncertain input and output data

In this paper, a robust support vector regression (RSVR) method with uncertain input and output data is studied. First, the data uncertainties are investigated under a stochastic framework and two linear robust formulations are derived. Linear formulations robust to ellipsoidal uncertainties are also considered from a geometric perspective. Second, kernelized RSVR formulations are established for nonlinear regression problems. Both linear and nonlinear formulations are converted to second-order cone programming problems, which can be solved efficiently by the interior point method. Simulation demonstrates that the proposed method outperforms existing RSVRs in the presence of both input and output data uncertainties.

Visual event recognition in videos by learning from Web data

We propose a visual event recognition framework for consumer videos by leveraging a large amount of loosely labeled web videos (e.g., from YouTube). Observing that consumer videos generally contain large intraclass variations within the same type of events, we first propose a new method, called Aligned Space-Time Pyramid Matching (ASTPM), to measure the distance between any two video clips. Second, we propose a new transfer learning method, referred to as Adaptive Multiple Kernel Learning (A-MKL), in order to 1) fuse the information from multiple pyramid levels and features (i.e., space-time features and static SIFT features) and 2) cope with the considerable variation in feature distributions between videos from two domains (i.e., web video domain and consumer video domain). For each pyramid level and each type of local features, we first train a set of SVM classifiers based on the combined training set from two domains by using multiple base kernels from different kernel types and parameters, which are then fused with equal weights to obtain a prelearned average classifier. In A-MKL, for each event class we learn an adapted target classifier based on multiple base kernels and the prelearned average classifiers from this event class or all the event classes by minimizing both the structural risk functional and the mismatch between data distributions of two domains. Extensive experiments demonstrate the effectiveness of our proposed framework that requires only a small number of labeled consumer videos by leveraging web data. We also conduct an in-depth investigation on various aspects of the proposed method A-MKL, such as the analysis on the combination coefficients on the prelearned classifiers, the convergence of the learning algorithm, and the performance variation by using different proportions of labeled consumer videos. Moreover, we show that A-MKL using the prelearned classifiers from all the event classes leads to better performance when compared with A-MK- using the prelearned classifiers only from each individual event class.