As an important machine learning topic, dimensionality reduction has been widely studied and utilized in various kinds of areas. A multitude of dimensionality reduction methods have been developed, among which unsupervised dimensionality reduction is more desirable when obtaining label information requires onerous work. However, most previous unsupervised dimensionality reduction methods call for an affinity graph constructed beforehand, with which the following dimensionality reduction steps can be then performed. Separation of graph construction and dimensionality reduction leads the dimensionality reduction process highly dependent on quality of the input graph. In this paper, we propose a novel graph embedding method for unsupervised dimensionality reduction. We simultaneously conduct dimensionality reduction along with graph construction by assigning adaptive and optimal neighbors according to the projected local distances. Our method doesn’t need an affinity graph constructed in advance, but instead learns the graph concurrently with dimensionality reduction. Thus, the learned graph is optimal for dimensionality reduction. Meanwhile, our learned graph has an explicit block diagonal structure, from which the clustering results could be directly revealed without any postprocessing steps. Extensive empirical results on dimensionality reduction as well as clustering are presented to corroborate the performance of our method.

In predicate invention (PI), new predicates are introduced into a logical theory, usually by rewriting a group of closely-related rules to use a common invented predicate as a "subroutine". PI is difficult, since a poorly-chosen invented predicate may lead to error cascades. Here we suggest a "soft" version of predicate invention: instead of explicitly creating new predicates, we implicitly group closely-related rules by using structured sparsity to regularize their parameters together. We show that soft PI, unlike hard PI, consistently improves over previous strong baselines for structure-learning on two large-scale tasks.

SVM in advance, and this limits their applications to largescale problems. To address this issue, several methods for Support vector machines (SVMs) play a very dominant selecting a set of basis vectors are proposed. They include role in data classification due to their good sampling from the training set in the Nystrom method generalization performance. However, they suffer [Williams and Seeger, 2001] and variants of the Incomplete from the high computational complexity in the Cholesky factorization [Bach and Jordan, 2005], core vector classification phase when there are a considerable machine (CVM) [Tsang et al., 2005], relevance vector machine number of support vectors (SVs). Then it is desirable (RVM)[Tipping, 2001], and relevance units machine to design efficient algorithms in the classification (RUM)[Gao and Zhang, 2009]. Wu et al. [Wu et al., 2006] phase to deal with the datasets of realtime add one constraint on the number of basis vectors to the standard pattern recognition systems. To this end, we SVM optimization problem, and then solve this modified propose a novel classifier called HMLSVMs (Hierarchical nonconvex problem to build sparse kernel learning algorithms Mixing Linear Support Vector Machines) (SKLA). Joachims and Yu [Joachims and Yu, 2009] in this paper, which has a hierarchical structure explore a new sparse kernel SVMs via cutting plane training, with a mixing linear SVMs classifier at each node called cutting-plane subspace pursuit (CPSP).Although and predicts the label of a sample using only a the above methods prunes the SVs and reduces computational few hyperplanes. We also give a generalization complexity in classification phase, when a new test sample is error bound for the class of locally linear SVMs introduced, they still need to compare it with these pruned (LLSVMs) based on the Rademacher theory, which SVs via kernel calculations to predict the label of the test ensures that overfitting can be effectively avoided.

Multimodal hashing is essential to cross-media similarity search for its low storage cost and fast query speed. Most existing multimodal hashing methods embedded heterogeneous data into a common low-dimensional Hamming space, and then rounded the continuous embeddings to obtain the binary codes. Yet they usually neglect the inherent discrete nature of hashing for relaxing the discrete constraints, which will cause degraded retrieval performance especially for long codes. For this purpose, a novel Semantic Topic Multimodal Hashing (STMH) is developed by considering latent semantic information in coding procedure. It first discovers clustering patterns of texts and robust factorizes the matrix of images to obtain multiple semantic topics of texts and concepts of images. Then the learned multimodal semantic features are transformed into a common subspace by their correlations. Finally, each bit of unified hash code can be generated directly by figuring out whether a topic or concept is contained in a text or an image. Therefore, the obtained model by STMH is more suitable for hashing scheme as it directly learns discrete hash codes in the coding process. Experimental results demonstrate that the proposed method outperforms several state-of-the-art methods.

We introduce a novel framework for option discovery and learning transfer in complex domains that are represented as object-oriented Markov decision processes (OO-MDPs) [Diuk et al., 2008]. Our framework, Portable Option Discovery (POD), extends existing option discovery methods, and enables transfer across related but different domains by providing an unsupervised method for finding a mapping between object-oriented domains with different state spaces. The framework also includes heuristic approaches for increasing the efficiency of the mapping process. We present the results of applying POD to Pickett and Barto's [2002] PolicyBlocks and MacGlashan's [2013] Option-Based Policy Transfer in two application domains. We show that our approach can discover options effectively, transfer options among different domains, and improve learning performance with low computational overhead.

Generating a coherent synopsis and revealing the development threads for news stories from the increasing amounts of news content remains aformidable challenge. In this paper, we proposed a hddCRP (hybird distant-dependent ChineseRestaurant Process) based HierARChical tOpic model for news Article cLustering, abbreviated as CHARCOAL. Given a bunch of news articles, the outcome of CHARCOAL is threefold: 1) it aggregates relevant new articles into clusters (i.e., stories); 2) it disentangles the chain links (i.e., storyline) between articles in their describing story; 3) it discerns the topics that each story is assigned (e.g., Malaysia Airlines Flight 370 story belongs to the aircraft accident topic and U.S presidential election stories belong to the politics topic). CHARCOAL completes this task by utilizing a hddCRP as prior, and the entities (e.g., names of persons, organizations, or locations) that appear in news articles as clues. Moveover, the adaptation of nonparametric nature in CHARCOAL makes our model can adaptively learn the appropriate number of stories and topics from news corpus. The experimental analysis and results demonstrate both interpretability and superiority of the proposed approach.

In the context of sequence processing, we study the relationship between single-layer feedforward neural networks,that have simultaneous access to all items composing a sequence, and single-layer recurrent neural networks which access information one step at a time.We treat both linear and nonlinear networks, describing a constructive procedure, based on linear autoencoders for sequences, that given a feedforward neural network shows how to define a recurrent neural network that implements the same function in time. Upper bounds on the required number of hidden units for the recurrent network as a function of some features of the feedforward network are given. By separating the functional from the memory component, the proposed procedure suggests new efficient learning as well as interpretation procedures for recurrent neural networks.

We propose a technique for decomposing and compressing the dataset in the parameter learning problem in Markov random fields. Our technique applies to incomplete datasets and exploits variables that are always observed in the given dataset. We show that our technique allows exact computation of the gradient and the likelihood, and can lead to orders-of-magnitude savings in learning time.

It is also useful for scientific discovery like in neuroscience, like correlation of X, Y only test for (univariate) to see if a stimulus X (say an image) is independent linear independence, natural alternatives like of the brain activity Y (say fMRI) in a relevant part of mutual information of X, Y are hard to estimate the brain. Since detecting nonlinear correlations is much easier due to a serious curse of dimensionality. A recent than estimating a nonparametric regression function (of approach, avoiding both issues, estimates norms of Y onto X), it can be done at smaller sample sizes, with further an operator in Reproducing Kernel Hilbert Spaces samples collected for estimation only if an effect is detected (RKHSs). Our main contribution is strong empirical by the hypothesis test. For such situations, correlation evidence that by employing shrunk operators only tests for univariate linear independence, while other when the sample size is small, one can attain an improvement statistics like mutual information that do characterize multivariate in power at low false positive rates. We independence are hard to estimate from data, suffering analyze the effects of Stein shrinkage on a popular from a serious curse of dimensionality. A recent popular test statistic called HSIC (Hilbert-Schmidt Independence approach for this problem (and a related two-sample testing Criterion). Our observations provide insights problem) involve the use of quantities defined in reproducing into two recently proposed shrinkage estimators, kernel Hilbert spaces (RKHSs) - see [Gretton et al., 2006; SCOSE and FCOSE - we prove that SCOSE Harchaoui et al., 2007; Gretton et al., 2005b; 2005a].

We introduce a novel kernel that upgrades the Weisfeiler-Lehman and other graph kernels to effectively exploit high-dimensional and continuous vertex attributes. Graphs are first decomposed into subgraphs. Vertices of the subgraphs are then compared by a kernel that combines the similarity of their labels and the similarity of their structural role, using a suitable vertex invariant. By changing this invariant we obtain a family of graph kernels which includes generalizations of Weisfeiler-Lehman, NSPDK, and propagation kernels. We demonstrate empirically that these kernels obtain state-of-the-art results on relational data sets.