Domain adaptation solves a learning problem in a target domain by utilizing the training data in a different but related source domain. Intuitively, discovering a good feature representation across domains is crucial. In this paper, we propose to find such a representation through a new learning method, transfer component analysis ( TCA ), for domain adaptation. TCA tries to learn some transfer components across domains in a Reproducing Kernel Hilbert Space (RKHS) using Maximum Mean Discrepancy (MMD). In the subspace spanned by these transfer components , data distributions in different domains are close to each other. As a result, with the new representations in this subspace, we can apply standard machine learning methods to train classifiers or regression models in the source domain for use in the target domain. The main contribution of our work is that we propose a novel feature representation in which to perform domain adaptation via a new parametric kernel using feature extraction methods, which can dramatically minimize the distance between domain distributions by projecting data onto the learned transfer components . Furthermore, our approach can handle large datsets and naturally lead to out-of-sample generalization. The effectiveness and efficiency of our approach in are verified by experiments on two real-world applications: cross-domain indoor WiFi localization and cross-domain text classification.

In this paper, we propose a new spectral clustering method, referred to as Spectral Embedded Clustering (SEC), to minimize the normalized cut criterion in spectral clustering as well as control the mismatch between the cluster assignment matrix and the low dimensional embedded representation of the data. SEC is based on the observation that the cluster assignment matrix of high dimensional data can be represented by a low dimensional linear mapping of data. We also discover the connection between SEC and other clustering methods, such as spectral clustering, Clustering with local and global regularization, K-means and Discriminative K-means. The experiments on many real-world data sets show that SEC significantly outperforms the existing spectral clustering methods as well as K-means clustering related methods.

There has been intense interest in hierarchical reinforcement learning as a way to make Markov decision process planning more tractable, but there has been relatively little work on autonomously learning the hierarchy, especially in continuous domains. In this paper we present a method for learning a hierarchy of actions in a continuous environment. Our approach is to learn a qualitative representation of the continuous environment and then to define actions to reach qualitative states. Our method learns one or more options to perform each action. Each option is learned by first learning a dynamic Bayesian network (DBN). We approach this problem from a developmental robotics perspective. The agent receives no extrinsic reward and has no external direction for what to learn. We evaluate our work using a simulation with realistic physics that consists of a robot playing with blocks at a table.

The web holds tremendous potential as a source of training data for visual classification. However, web images must be correctly indexed and labeled before this potential can be realized. Accordingly, there has been considerable recent interest in collecting imagery from the web using image search engines to build databases for object and scene recognition research. While search engines can provide rough sets of image data, results are noisy and this leads to problems when training classifiers. In this paper we propose a semi-supervised model for automatically collecting clean example imagery from the web. Our approach includes both visual and textual web data in a unified framework. Minimal supervision is enabled by the selective use of generative and discriminative elements in a probabilistic model and a novel learning algorithm. We show through experiments that our model discovers good training images from the web with minimal manual work. Classifiers trained using our method significantly outperform analogous baseline approaches on the Caltech-256 dataset.

A central goal of transfer learning is to enable learning when training data from the domain of interest is limited. Yet, work on transfer across relational domains has so far focused on the case where there is a significant amount of target data. This paper bridges this gap by studying transfer when the amount of target data is minimal and consists of information about just a handful of entities. In the extreme case, only a single entity is known. We present the SR2LR algorithm that finds an effective mapping of predicates from a source model to the target domain in this setting and thus renders pre-existing knowledge useful to the target task. We demonstrate SR2LR's effectiveness in three benchmark relational domains on social interactions and study its behavior as information about an increasing number of entities becomes available.

Boltzmann Machines are a powerful class of undirected graphical models. Originally proposed as artificial neural networks, they can be regarded as a type of Markov Random Field in which the connection weights between nodes are symmetric and learned from data. They are also closely related to recent models such as Markov logic networks and Conditional Random Fields. A major challenge for Boltzmann machines (as well as other graphical models) is speeding up learning for large-scale problems. The heart of the problem lies in efficiently and effectively approximating the partition function. In this paper, we propose a new efficient learning algorithm for Boltzmann machines that allows them to be applied to problems with large numbers of random variables. We introduce a new large-margin variational approximation to the partition function that allows Boltzmann machines to be trained using a support vector machine (SVM) style learning algorithm. For discriminative learning tasks, these large margin Boltzmann machines provide an alternative approach to structural SVMs. We show that these machines have low sample complexity and derive a generalization bound. Our results demonstrate that on multi-label classification problems, large margin Boltzmann machines achieve orders of magnitude faster performance than structural SVMs and also outperform structural SVMs on problems with large numbers of labels.

Typical graph-theoretic approaches for semi-supervised classification infer labels of unlabeled instances with the help of graph Laplacians. Founded on the spectral decomposition of the graph Laplacian, this paper learns a kernel matrix via minimizing the leave-one-out classification error on the labeled instances.  To this end, an efficient algorithm is presented based on linear programming, resulting in a transductive spectral kernel. The idea of our algorithm stems from regularization methodology and also has a nice interpretation in terms of spectral clustering. A simple classifier can be readily built upon the learned kernel, which suffices to give prediction for any data point aside from those in the available dataset. Besides this usage, the spectral kernel can be effectively used in tandem with conventional kernel machines such as SVMs. We demonstrate the efficacy of the proposed algorithm through experiments carried out on challenging classification tasks.

Kernel methods have been applied successfully in many applications. The kernel matrix plays an important role in kernel-based learning methods, but the ideal kernel matrix is usually unknown in practice and needs to be estimated. In this paper, we propose to directly learn the ideal kernel matrix (called the optimal neighborhood kernel matrix) from a pre-specified kernel matrix for improved classification performance. We assume that the pre-specified kernel matrix generated from the specific application is a noisy observation of the ideal one. The resulting optimal neighborhood kernel matrix is shown to be the summation of the pre-specified kernel matrix and a rank-one matrix. We formulate the problem of learning the optimal neighborhood kernel as a constrained quartic problem, and propose to solve it using two methods: level method and constrained gradient descent. Empirical results on several benchmark data sets demonstrate the efficiency and effectiveness of the proposed algorithms.

Recognition of chatting activities in social interactions is useful for constructing human social networks. However, the existence of multiple people involved in multiple dialogues presents special challenges. To model the conversational dynamics of concurrent chatting behaviors, this paper advocates Factorial Conditional Random Fields (FCRFs) as a model to accommodate co-temporal relationships among multiple activity states. In addition, to avoid the use of inefficient Loopy Belief Propagation (LBP) algorithm, we propose using Iterative Classification Algorithm (ICA) as the inference method for FCRFs. We designed experiments to compare our FCRFs model with two dynamic probabilistic models, Parallel Condition Random Fields (PCRFs) and Hidden Markov Models (HMMs), in learning and decoding based on auditory data. The experimental results show that FCRFs outperform PCRFs and HMM-like models. We also discover that FCRFs using the ICA inference approach not only improves the recognition accuracy but also takes significantly less time than the LBP inference method.

Object recognition using image-set or video sequence as input tends to be more robust since image-set or video sequence provides much more information than single snap-shot about the variability in the appearance of the target subject. Constrained Mutual Subspace Method (CMSM) is one of the state-of-the-art algorithms for imageset based object recognition by first projecting the image-set patterns onto the so-called generalized difference subspace then classifying based on the principal angle based mutual subspace distance. By treating the subspace bases for each image-set patterns as basic elements in the grassmann manifold, this paper presents a framework for robust image-set based recognition by CMSM based ensemble learning in a boosting way. The proposed Boosting Constrained Mutual Subspace Method(BCMSM) improves the original CMSM in the following ways: a) The proposed BCMSM algorithm is insensitive to the dimension of the generalized differnce subspace while the performance of the original CMSM algorithm is quite dependent on the dimension and the selecting of optimum choice is quite empirical and case-dependent; b) By taking advantage of both boosting and CMSM techniques, the generalization ability is improved and much higher classification performance can be achieved. Extensive experiments on real-life data sets (two face recognition tasks and one 3D object category classification task) show that the proposed method outperforms the previous state-of-the-art algorithms greatly in terms of classification accuracy.