Locality%2Dconstrained Linear Coding for image classification.
The traditional SPM approach based on bag-of-features (BoF) requires nonlinear classifiers to achieve good image classification performance. This paper presents a simple but effective coding scheme called Locality-constrained Linear Coding (LLC) in place of the VQ coding in traditional SPM. LLC utilizes the locality constraints to project each descriptor into its local-coordinate system, and the projected coordinates are integrated by max pooling to generate the final representation. With linear classifier, the proposed approach performs remarkably better than the traditional nonlinear SPM, achieving state-of-the-art performance on several benchmarks. Compared with the sparse coding strategy [22], the objective function used by LLC has an analytical solution. In addition, the paper proposes a fast approximated LLC method by first performing a K-nearest-neighbor search and then solving a constrained least square fitting problem, bearing computational complexity of O(M + K2). Hence even with very large codebooks, our system can still process multiple frames per second. This efficiency significantly adds to the practical values of LLC for real applications. 1.
Supervised translation%2Dinvariant sparse coding.

Efficient Highly Over%2DComplete Sparse Coding Using a Mixture Model.
Abstract. Sparse coding of sensory data has recently attracted notable attention in research of learning useful features from the unlabeled data. Empirical studies show that mapping the data into a significantly higherdimensional space with sparse coding can lead to superior classification performance. However, computationally it is challenging to learn a set of highly over-complete dictionary bases and to encode the test data with the learned bases. In this paper, we describe a mixture sparse coding model that can produce high-dimensional sparse representations very efficiently. Besides the computational advantage, the model effectively encourages data that are similar to each other to enjoy similar sparse representations. What’s more, the proposed model can be regarded as an approximation to the recently proposed local coordinate coding (LCC), which states that sparse coding can approximately learn the nonlinear manifold of the sensory data in a locally linear manner. Therefore, the feature learned by the mixture sparse coding model works pretty well with linear classifiers. We apply the proposed model to PASCAL VOC 2007 and 2009 datasets for the classification task, both achieving stateof-the-art performances. Key words: Sparse coding, highly over-complete dictionary training, mixture model, mixture sparse coding, image classification, PASCAL VOC challenge 1


Predicting Facial Beauty without Landmarks.
Abstract. A fundamental task in artificial intelligence and computer vision is to build machines that can behave like a human in recognizing a broad range of visual concepts. This paper aims to investigate and develop intelligent systems for learning the concept of female facial beauty and producing human-like predictors. Artists and social scientists have long been fascinated by the notion of facial beauty, but study by computer scientists has only begun in the last few years. Our work is notably different from and goes beyond previous works in several aspects: 1) we focus on fully-automatic learning approaches that do not require costly manual annotation of landmark facial features but simply take the raw pixels as inputs; 2) our study is based on a collection of data that is an order of magnitude larger than that of any previous study; 3) we imposed no restrictions in terms of pose, lighting, background, expression, age, and ethnicity on the face images used for training and testing. These factors significantly increased the difficulty of the learning task. We show that a biologically-inspired model with multiple layers of trainable feature extractors can produce results that are much more human-like than the previously used eigenface approach. Finally, we develop a novel visualization method to interpret the learned model and revealed the existence of several beautiful features that go beyond the current averageness and symmetry hypotheses. 1
Word%2Dlevel emphasis modelling in HMM%2Dbased speech synthesis.
Expressive speech synthesis has recently attracted great interest. Word-level emphasis is an important form of expressiveness to distinguish between what is the focus of the utterance, and what the computer system expects to be known by the user. Previous work on emphasis synthesis requires emphatic data collected specifically for that task. In this paper, a statistical approach that models and extracts word-level emphasis patterns from natural speech is investigated within the HMM based speech synthesis framework. Compared to emphatic speech collected specifically for this task, the cues of emphasis in natural speech are weaker and heavily affected by various suprasegmental features. Two new decision tree clustering approaches, two-pass and factorized decision tree, are proposed to effectively address this problem. Experiments show that both approaches can convey emphasis significantly better than traditional decision tree clustering and HMM adaptation. While the two-pass decision tree approach outperformed the factorized decision tree approach in an emphasis synthesis test, the latter led to significantly better naturalness and hence achieved a better overall balance. Index Terms — HMM based synthesis, expressive speech synthesis, decision tree 1.
Improved Local Coordinate Coding using Local Tangents.
Local Coordinate Coding (LCC), introduced in (Yu et al., 2009), is a high dimensional nonlinear learning method that explicitly takes advantage of the geometric structure of the data. Its successful use in the winning system of last year’s Pascal image classification Challenge (Everingham, 2009) shows that the ability to integrate geometric information is critical for some real world machine learning applications. This paper further develops the idea of integrating geometry in machine learning by extending the original LCC method to include local tangent directions. These new correction terms lead to better approximation of high dimensional nonlinear functions when the underlying data manifold is locally relatively flat. The method significantly reduces the number of anchor points needed in LCC, which not only reduces computational cost, but also improves prediction performance. Experiments are included to demonstrate that this method is more effective than the original LCC method on some image classification tasks. 1.
3D Convolutional Neural Networks for Human Action Recognition.
We consider the fully automated recognition of actions in uncontrolled environment. Most existing work relies on domain knowledge to construct complex handcrafted features from inputs. In addition, the environments are usually assumed to be controlled. Convolutional neural networks (CNNs) are a type of deep models that can act directly on the raw inputs, thus automating the process of feature construction. However, such models are currently limited to handle 2D inputs. In this paper, we develop a novel 3D CNN model for action recognition. This model extracts features from both spatial and temporal dimensions by performing 3D convolutions, thereby capturing the motion information encoded in multiple adjacent frames. The developed model generates multiple channels of information from the input frames, and the final feature representation is obtained by combining information from all channels. We apply the developed model to recognize human actions in real-world environment, and it achieves superior performance without relying on handcrafted features. 1.
Chinese calligraphy specific style rendering system.

The Hidden Information State model A practical framework for POMDP%2Dbased spoken dialogue management.
Computer Speech and Language xxx (2009) xxx–xxx COMPUTER SPEECH AND LANGUAGE www.elsevier.com/locate/csl
Feature Selection for Gene Expression Using Model%2DBased Entropy.
Abstract—Gene expression data usually contain a large number of genes, but a small number of samples. Feature selection for gene expression data aims at finding a set of genes that best discriminate biological samples of different types. Using machine learning techniques, traditional gene selection based on empirical mutual information suffers the data sparseness issue due to the small number of samples. To overcome the sparseness issue, we propose a model-based approach to estimate the entropy of class variables on the model, instead of on the data themselves. Here, we use multivariate normal distributions to fit the data, because multivariate normal distributions have maximum entropy among all real-valued distributions with specified mean and standard deviation, and are widely used to approximate various distributions. Given that the data follow a multivariate normal distribution, since the conditional distribution of class variables given the selected features is normal distribution, its entropy can be computed with the log-determinant of its covariance matrix. Because of the large number of genes, the computation of all possible log-determinants is not efficient. We propose several algorithms to largely reduce the computational cost. The experiments on seven gene datasets and the comparison with other five approaches show the accuracy of the multivariate Gaussian generative model for feature selection, and the efficiency of our algorithms.
Unsupervised training and directed manual transcription for LVCSR.


Performance Analysis of Large CDMA Random Access Systems with Retransmission Diversity over Fading Channels.


Program Sifting Select Property%2DRelated Functions for Language%2DBased Static Analysis.

Lazy Symbolic Evaluation and Its Path Constraints Solution.

Linear spatial pyramid matching using sparse coding for image classification.
Recently SVMs using spatial pyramid matching (SPM) kernel have been highly successful in image classification. Despite its popularity, these nonlinear SVMs have a complexity O(n 2 ∼ n 3) in training and O(n) in testing, where n is the training size, implying that it is nontrivial to scaleup the algorithms to handle more than thousands of training images. In this paper we develop an extension of the SPM method, by generalizing vector quantization to sparse coding followed by multi-scale spatial max pooling, and propose a linear SPM kernel based on SIFT sparse codes. This new approach remarkably reduces the complexity of SVMs to O(n) in training and a constant in testing. In a number of image categorization experiments, we find that, in terms of classification accuracy, the suggested linear SPM based on sparse coding of SIFT descriptors always significantly outperforms the linear SPM kernel on histograms, and is even better than the nonlinear SPM kernels, leading to state-of-the-art performance on several benchmarks by using a single type of descriptors. 1.
Probablistic modelling of F0 in unvoiced regions in HMM based speech synthesis.
HMM based synthesis has attracted great interest due to its compact and flexible modelling of spectral and prosodic parameters. In this approach, short term spectra, fundamental frequency (F0) and duration are simultaneously modelled by multi-stream HMMs. However, since F0 values in unvoiced regions are normally considered as undefined, it is difficult to use standard HMMs for F0 modelling. The currently preferred solution to this is to use a multi-space distribution HMM (MSDHMM) in which discrete distributions are used for modelling the voiced/unvoiced decision and continuous Gaussian distributions are used for modelling the F0 values within the voiced regions. However, the assumption of undefined unvoiced F0 regions and the special structure of the MSDHMM lead to limitations in the accurate modelling of F0 patterns. In this paper an alternative is explored whereby unvoiced F0 values are assumed to exist and are modelled within the standard HMM framework using a globally tied distribution (GTD). Subjective evaluations show that these regular HMMs with GTD can produce significant improvements in the naturalness of the synthesised speech compared to the MSDHMM, and furthermore, the method is insensitive to the exact method used for unvoiced F0 generation. Index Terms — HMM based synthesis, F0 modelling
Spoken language understanding from unaligned data using discriminative classification models.
While data-driven methods for spoken language understanding reduce maintenance and portability costs compared with handcrafted parsers, the collection of word-level semantic annotations for training remains a time-consuming task. A recent line of research has focused on building generative models from unaligned semantic representations, using expectation-maximisation techniques to align semantic concepts. This paper presents an efficient, simple technique that parses a semantic tree by recursively calling discriminative semantic classification models. Results show that it outperforms existing generative models, while performance is close to more complex grammar induction techniques. We also show that our method is robust to speech recognition errors, by improving over a handcrafted parser previously used for dialogue data collection. Index Terms — semantic analysis, spoken dialogue systems, spoken language understanding
Large%2Dscale collaborative prediction using a nonparametric random effects model.
A nonparametric model is introduced that allows multiple related regression tasks to take inputs from a common data space. Traditional transfer learning models can be inappropriate if the dependence among the outputs cannot be fully resolved by known inputspecific and task-specific predictors. The proposed model treats such output responses as conditionally independent, given known predictors and appropriate unobserved random effects. The model is nonparametric in the sense that the dimensionality of random effects is not specified a priori but is instead determined from data. An approach to estimating the model is presented uses an EM algorithm that is efficient on a very large scale collaborative prediction problem. The obtained prediction accuracy is competitive with state-of-the-art results. 1.
Tutorial summary Learning with dependencies between several response variables.

Style%2Dconsistency calligraphy synthesis system in digital library.

Style%2Dconsistency calligraphy synthesis system in digital library.

k%2DNearest Neighbor Monte%2DCarlo Control Algorithm for POMDP%2DBased Dialogue Systems.
In real-world applications, modelling dialogue as a POMDP requires the use of a summary space for the dialogue state representation to ensure tractability. Sub-optimal estimation of the value function governing the selection of system responses can then be obtained using a grid-based approach on the belief space. In this work, the Monte-Carlo control technique is extended so as to reduce training over-fitting and to improve robustness to semantic noise in the user input. This technique uses a database of belief vector prototypes to choose the optimal system action. A locally weighted k-nearest neighbor scheme is introduced to smooth the decision process by interpolating the value function, resulting in higher user simulation performance. 1
Fast nonparametric matrix factorization for large%2Dscale collaborative filtering.
With the sheer growth of online user data, it becomes challenging to develop preference learning algorithms that are sufficiently flexible in modeling but also affordable in computation. In this paper we develop nonparametric matrix factorization methods by allowing the latent factors of two low-rank matrix factorization methods, the singular value decomposition (SVD) and probabilistic principal component analysis (pPCA), to be data-driven, with the dimensionality increasing with data size. We show that the formulations of the two nonparametric models are very similar, and their optimizations share similar procedures. Compared to traditional parametric low-rank methods, nonparametric models are appealing for their flexibility in modeling complex data dependencies. However, this modeling advantage comes at a computational price — it is highly challenging to scale them to large-scale problems, hampering their application to applications such as collaborative filtering. In this paper we introduce novel optimization algorithms, which are simple to implement, which allow learning both nonparametric matrix factorization models to be highly efficient on large-scale problems. Our experiments on EachMovie and Netflix, the two largest public benchmarks to date, demonstrate that the nonparametric models make more accurate predictions of user ratings, and are computationally comparable or sometimes even faster in training, in comparison with previous state-of-the-art parametric matrix factorization models.
Performance of large CDMA random access systems with retransmission diversity over fading channels.

Relay%2Daided multiple antenna transmissions for wireless backhaul applications.



High data rate relay transmissions with multiple antennas.

Training Hierarchical Feed%2DForward Visual Recognition Models Using Transfer Learning from Pseudo%2DTasks.
Abstract. Building visual recognition models that adapt across different domains is a challenging task for computer vision. While feature-learning machines in the form of hierarchial feed-forward models (e.g., convolutional neural networks) showed promise in this direction, they are still difficult to train especially when few training examples are available. In this paper, we present a framework for training hierarchical feed-forward models for visual recognition, using transfer learning from pseudo tasks. These pseudo tasks are automatically constructed from data without supervision and comprise a set of simple pattern-matching operations. We show that these pseudo tasks induce an informative inverse-Wishart prior on the functional behavior of the network, offering an effective way to incorporate useful prior knowledge into the network training. In addition to being extremely simple to implement, and adaptable across different domains with little or no extra tuning, our approach achieves promising results on challenging visual recognition tasks, including object recognition, gender recognition, and ethnicity recognition. 1
Visual Knowledge Reasoning on Typed Categorical Structure.

Unsupervised discriminative adaptation using discriminative mapping transforms.
The most commonly used approaches to speaker adaptation are based on linear transforms, as these can be robustly estimated using limited adaptation data. Although significant gains can be obtained using discriminative criteria for training acoustic models, maximum likelihood (ML) estimated transforms are used for unsupervised adaptation. This is because discriminatively trained transforms are highly sensitive to errors in the adaptation hypothesis. This paper describes a new framework for estimating transforms that are discriminative in nature, but are less sensitive to this hypothesis issue. A discriminative, speaker-independent, mapping transformation is estimated during training. This transform is obtained after a speaker-specific ML-estimated transform has been applied. During recognition an ML speaker-specific transform is found and the speaker-independent discriminative mapping transform then applied. This allows a transform which is discriminative in nature to be indirectly estimated, whilst only requiring an ML speaker-specific transform to be found during recognition. The scheme is evaluated on an English conversational telephone speech task, where it significantly outperforms both standard ML and discriminatively trained transforms. Index Terms — Speaker adaptation, discriminative training. 1.
Towards a Categorical Matching Method to Process High%2DDimensional Emergency Knowledge Structures.

Toward a Categorical Link Grammar for Knowledge Piece Generation in Emergency Decision%2DMaking Support.

Deep Learning with Kernel Regularization for Visual Recognition.
In this paper we aim to train deep neural networks for rapid visual recognition. The task is highly challenging, largely due to the lack of a meaningful regularizer on the functions realized by the networks. We propose a novel regularization method that takes advantage of kernel methods, where a given kernel function represents prior knowledge about the recognition task of interest. We derive an efficient algorithm using stochastic gradient descent, and demonstrate encouraging results on a wide range of recognition tasks, in terms of both accuracy and speed. 1

Stochastic Relational Models for Large%2Dscale Dyadic Data using MCMC.
Stochastic relational models (SRMs) [15] provide a rich family of choices for learning and predicting dyadic data between two sets of entities. The models generalize matrix factorization to a supervised learning problem that utilizes attributes of entities in a hierarchical Bayesian framework. Previously variational Bayes inference was applied for SRMs, which is, however, not scalable when the size of either entity set grows to tens of thousands. In this paper, we introduce a Markov chain Monte Carlo (MCMC) algorithm for equivalent models of SRMs in order to scale the computation to very large dyadic data sets. Both superior scalability and predictive accuracy are demonstrated on a collaborative filtering problem, which involves tens of thousands users and half million items. 1 Stochastic Relational Models Stochastic relational models (SRMs) [15] are generalizations of Gaussian process (GP) models [11] to the relational domain, where each observation is a dyadic datum, indexed by a pair of entities. They model dyadic data by a multiplicative interaction of two Gaussian process priors.

Combining MIMO and relaying gains for highly efficient wireless backhaul.

trNon%2Dgreedy active learning for text categorization using convex ansductive experimental design.

Surveillance Event Detection.
We have developed and evaluated three generalized systems for event detection. The first system is a simple brute force search method, where each space-time location in the video is evaluated by a binary decision rule on whether it contains the event or not. The second system is build on top of a head tracker to avoid costly brute force searching. The decision stage is a combination of state of the art feature extractors and classifiers. Our third system has a probabilistic framework. From the observations, the pose of the people are estimated and used to determine the presence of event. Finally we introduce two ad-hoc methods that were designed to specifically detect OpposingFlow and TakePicture events. The results are promising as we are able to get good results on several event categories, while for all events we have gained valuable insights and experience. 1
Learning multiple graphs for document recommendations.
The Web offers rich relational data with different semantics. In this paper, we address the problem of document recommendation in a digital library, where the documents in question are networked by citations and are associated with other entities by various relations. Due to the sparsity of a single graph and noise in graph construction, we propose a new method for combining multiple graphs to measure document similarities, where different factorization strategies are used based on the nature of different graphs. In particular, the new method seeks a single low-dimensional embedding of documents that captures their relative similarities in a latent space. Based on the obtained embedding, a new recommendation framework is developed using semisupervised learning on graphs. In addition, we address the scalability issue and propose an incremental algorithm. The new incremental method significantly improves the efficiency by calculating the embedding for new incoming documents only. The new batch and incremental methods are evaluated on two real world datasets prepared from CiteSeer. Experiments demonstrate significant quality improvement for our batch method and significant efficiency improvement with tolerable quality loss for our incremental method.
A 1000%2Dword vocabulary
The Hearsay-II system, developed during the DARPA-sponsored five-year speech-understanding research program, represents both a specific solution to the speech-understanding problem and a general framework for coordinating independent processes to achieve cooperative problem-solving behavior. As a computational problem, speech understanding reflects a large number of intrinsically interesting issues. Spoken sounds are achieved by a long chain of successive transformations, from intentions, through semantm and syntactic structurmg, to the eventually resulting audible acoustic waves. As a consequence, interpreting speech means effectively inverting these transformations to recover the speaker&#039;s intention from the sound. At each step in the interpretive process, ambiguity and uncertainty arise. The Hearsay-II problem-solving framework reconstructs an intention from hypothetmal interpretations formulated at various levels of abstraction. In additmn, it allocates hmlted processing resources fwst to the most promising incremental actions. The final configuration of the Hearsay-II system comprises problem-solving components to generate and evaluate speech hypotheses, and a focus-of-control mechanism to identify
Local learning projections.
This paper presents a Local Learning Projection (LLP) approach for linear dimensionality reduction. We first point out that the well known Principal Component Analysis (PCA) essentially seeks the projection that has the minimal global estimation error. Then we propose a dimensionality reduction algorithm that leads to the projection with the minimal local estimation error, and elucidate its advantages for classification tasks. We also indicate that LLP keeps the local information in the sense that the projection value of each point can be well estimated based on its neighbors and their projection values. Experimental results are provided to validate the effectiveness of the proposed algorithm. 1.
Robust multi%2Dtask learning with t%2Dprocesses.
Most current multi-task learning frameworks ignore the robustness issue, which means that the presence of “outlier ” tasks may greatly reduce overall system performance. We introduce a robust framework for Bayesian multitask learning, t-processes (TP), which are a generalization of Gaussian processes (GP) for multi-task learning. TP allows the system to effectively distinguish good tasks from noisy or outlier tasks. Experiments show that TP not only improves overall system performance, but can also serve as an indicator for the “informativeness ” of different tasks. 1.
Fast Inference in Infinite Hidden Relational Models.


Gaussian Process Models for Link Analysis and Transfer Learning.


Combining content and link for classification using matrix factorization.
The world wide web contains rich textual contents that are interconnected via complex hyperlinks. This huge database violates the assumption held by most of conventional statistical methods that each web page is considered as an independent and identical sample. It is thus difficult to apply traditional mining or learning methods for solving web mining problems, e.g., web page classification, by exploiting both the content and the link structure. The research in this direction has recently received considerable attention but are still in an early stage. Though a few methods exploit both the link structure or the content information, some of them combine the only authority information with the content information, and the others first decompose the link structure into hub and authority features, then apply them as additional document features. Being practically attractive for its great simplicity, this paper aims to design an algorithm that exploits both the content and linkage information, by carrying out a joint factorization on both the linkage adjacency matrix and the document-term matrix, and derives a new representation for web pages in a low-dimensional factor space, without explicitly separating them as content, hub or authority factors. Further analysis can be performed based on the compact representation of web pages. In the experiments, the proposed method is compared with state-of-the-art methods and demonstrates an excellent accuracy in hypertext classification on the WebKB and Cora benchmarks.
Bayesian Adaptive Inference and Adaptive Training.
Abstract—Large-vocabulary speech recognition systems are often built using found data, such as broadcast news. In contrast to carefully collected data, found data normally contains multiple acoustic conditions, such as speaker or environmental noise. Adaptive training is a powerful approach to build systems on such data. Here, transforms are used to represent the different acoustic conditions, and then a canonical model is trained given this set of transforms. This paper describes a Bayesian framework for adaptive training and inference. This framework addresses some limitations of standard maximum-likelihood approaches. In contrast to the standard approach, the adaptively trained system can be directly used in unsupervised inference, rather than having to rely on initial hypotheses being present. In addition, for limited adaptation data, robust recognition performance can be obtained. The limited data problem often occurs in testing as there is no control over the amount of the adaptation data available. In contrast, for adaptive training, it is possible to control the system complexity to reflect the available data. Thus, the standard point estimates may be used. As the integral associated with Bayesian adaptive inference is intractable, various marginalization approximations are described, including a variational Bayes approximation. Both batch and incremental modes of adaptive inference are discussed. These approaches are applied to adaptive training of maximum-likelihood linear regression and evaluated on a large-vocabulary speech recognition task. Bayesian adaptive inference is shown to significantly outperform standard approaches. Index Terms—Adaptive training, Bayesian adaptation, Bayesian inference, incremental, variational Bayes.
Variational Bayesian Dirichlet%2DMultinomial Allocation for Exponential Family Mixtures.
Abstract. This paper studies a Bayesian framework for density modeling with mixture of exponential family distributions. Variational Bayesian Dirichlet-Multinomial allocation (VBDMA) is introduced, which performs inference and learning efficiently using variational Bayesian methods and performs automatic model selection. The model is closely related to Dirichlet process mixture models and demonstrates similar automatic model selection in the variational Bayesian context. 1
Active learning via transductive experimental design.
This paper considers the problem of selecting the most informative experiments x to get measurements y for learning a regression model y = f(x). We propose a novel and simple concept for active learning, transductive experimental design, that explores available unmeasured experiments (i.e.,unlabeled data) and has a better scalability in comparison with classic experimental design methods. Our in-depth analysis shows that the new method tends to favor experiments that are on the one side hardto-predict and on the other side representative for the rest of the experiments. Efficient optimization of the new design problem is achieved through alternating optimization and sequential greedy search. Extensive experimental results on synthetic problems and three real-world tasks, including questionnaire design for preference learning, active learning for text categorization, and spatial sensor placement, highlight the advantages of the proposed approaches. 1.
Collaborative ordinal regression.
Ordinal regression has become an effective way of learning user preferences, but most research focuses on single regression problems. In this paper we introduce collaborative ordinal regression, where multiple ordinal regression tasks are handled simultaneously. Rather than modeling each task individually, we explore the dependency between ranking functions through a hierarchical Bayesian model and assign a common Gaussian Process (GP) prior to all individual functions. Empirical studies show that our collaborative model outperforms the individual counterpart in preference learning applications. 1.
Supervised probabilistic principal component analysis.
Principal component analysis (PCA) has been extensively applied in data mining, pattern recognition and information retrieval for unsupervised dimensionality reduction. When labels of data are available, e.g., in a classification or regression task, PCA is however not able to use this information. The problem is more interesting if only part of the input data are labeled, i.e., in a semi-supervised setting. In this paper we propose a supervised PCA model called SPPCA and a semi-supervised PCA model called S 2 PPCA, both of which are extensions of a probabilistic PCA model. The proposed models are able to incorporate the label information into the projection phase, and can naturally handle multiple outputs (i.e., in multi-task learning problems). We derive an efficient EM learning algorithm for both models, and also provide theoretical justifications of the model behaviors. SPPCA and S 2 PPCA are compared with other supervised projection methods on various learning tasks, and show not only promising performance but also good scalability.
Stochastic Relational Models for Discriminative Link Prediction.
We introduce a Gaussian process (GP) framework, stochastic relational models (SRM), for learning social, physical, and other relational phenomena where interactions between entities are observed. The key idea is to model the stochastic structure of entity relationships (i.e., links) via a tensor interaction of multiple GPs, each defined on one type of entities. These models in fact define a set of nonparametric priors on infinite dimensional tensor matrices, where each element represents a relationship between a tuple of entities. By maximizing the marginalized likelihood, information is exchanged between the participating GPs through the entire relational network, so that the dependency structure of links is messaged to the dependency of entities, reflected by the adapted GP kernels. The framework offers a discriminative approach to link prediction, namely, predicting the existences, strengths, or types of relationships based on the partially observed linkage network as well as the attributes of entities (if given). We discuss properties and variants of SRM and derive an efficient learning algorithm. Very encouraging experimental results are achieved on a toy problem and a user-movie preference link prediction task. In the end we discuss extensions of SRM to general relational learning tasks. 1
Infinite Hidden Relational Models.

Multi%2DOutput Regularized Feature Projection.

Discriminative cluster adaptive training.
Abstract—Multiple-cluster schemes, such as cluster adaptive training (CAT) or eigenvoice systems, are a popular approach for rapid speaker and environment adaptation. Interpolation weights are used to transform a multiple-cluster, canonical, model to a standard hidden Markov model (HMM) set representative of an individual speaker or acoustic environment. Maximum likelihood training for CAT has previously been investigated. However, in state-of-the-art large vocabulary continuous speech recognition systems, discriminative training is commonly employed. This paper investigates applying discriminative training to multiple-cluster systems. In particular, minimum phone error (MPE) update formulae for CAT systems are derived. In order to use MPE in this case, modifications to the standard MPE smoothing function and the prior distribution associated with MPE training are required. A more complex adaptive training scheme combining both interpolation weights and linear transforms, a structured transform (ST), is also discussed within the MPE training framework. Discriminatively trained CAT and ST systems were evaluated on a state-of-the-art conversational telephone speech task. These multiple-cluster systems were found to outperform both standard and adaptively trained systems. Index Terms—Cluster adaptive training (CAT), discriminative training, eigenvoices, minimum phone error (MPE), multiple-cluster HMM. I.
Multi%2DOutput Regularized Projection.

Hierarchy%2DRegularized Latent Semantic Indexing.

Dirichlet enhanced relational learning.
We apply nonparametric hierarchical Bayesian modelling to relational learning. In a hierarchical Bayesian approach, model parameters can be “personalized”, i.e., owned by entities or relationships, and are coupled via a common prior distribution. Flexibility is added in a nonparametric hierarchical Bayesian approach, such that the learned knowledge can be truthfully represented. We apply our approach to a medical domain where we form a nonparametric hierarchical Bayesian model for relations involving hospitals, patients, procedures and diagnosis. The experiments show that the additional flexibility in a nonparametric hierarchical Bayes approach results in a more accurate model of the dependencies between procedures and diagnosis and gives significantly improved estimates of the probabilities of future procedures. 1.
Learning Gaussian processes from multiple tasks.
We consider the problem of multi-task learning, that is, learning multiple related functions. Our approach is based on a hierarchical Bayesian framework, that exploits the equivalence between parametric linear models and nonparametric Gaussian processes (GPs). The resulting models can be learned easily via an EM-algorithm. Empirical studies on multi-label text categorization suggest that the presented models allow accurate solutions of these multi-task problems. 1.
Soft Clustering on Graphs.
We propose a simple clustering framework on graphs that encode pairwise data similarities. Unlike usual similarity-based methods, the approach softly assigns data to clusters in a probabilistic way. More importantly, a hierarchical clustering is naturally derived in this framework to gradually merge lower-level clusters into higher-level ones. A random walk analysis indicates that the algorithm exposes clustering structures in various resolutions, i.e., a higher level statistically models a longer-term diffusion on graphs and thus discovers a more global clustering structure. Finally we provide very encouraging experimental results. 1
A Probabilistic Clustering%2DProjection Model for Discrete Data.
Abstract. For discrete co-occurrence data like documents and words, calculating optimal projections and clustering are two different but related tasks. The goal of projection is to find a low-dimensional latent space for words, and clustering aims at grouping documents based on their feature representations. In general projection and clustering are studied independently, but they both represent the intrinsic structure of data and should reinforce each other. In this paper we introduce a probabilistic clustering-projection (PCP) model for discrete data, where they are both represented in a unified framework. Clustering is seen to be performed in the projected space, and projection explicitly considers clustering structure. Iterating the two operations turns out to be exactly the variational EM algorithm under Bayesian model inference, and thus is guaranteed to improve the data likelihood. The model is evaluated on two text data sets, both showing very encouraging results. 1
Multi%2Dlabel informed latent semantic indexing.
Canonical Correlation Analysis (CCA) is a useful technique for modeling dependencies between two (or more) sets of variables. Building upon the recently suggested probabilistic interpretation of CCA, we propose a nonparametric, fully Bayesian framework that can automatically select the number of correlation components, and effectively capture the sparsity underlying the projections. In addition, given (partially) labeled data, our algorithm can also be used as a (semi)supervised dimensionality reduction technique, and can be applied to learn useful predictive features in the context of learning a set of related tasks. Experimental results demonstrate the efficacy of the proposed approach for both CCA as a stand-alone problem, and when applied to multi-label prediction. 1
MIMO channel capacity and modeling issues on a measured indoor radio channel at 5.8 GHz.
In recent years, multiple-input multiple-output (MIMO) systems appear to be very promising since they can provide high data rates in environments with sucient scattering by exploiting the spatial domain. To design a real MIMO wireless system and predict its performance under certain circumstances, it is necessary to have accurate MIMO wireless channel models for dierent scenarios. This thesis presents dierent models for indoor MIMO radio propagation channels based on 5.2 GHz indoor MIMO channel measurements.
Heterogenous Data Fusion via a Probabilistic Latent%2DVariable Model.
Abstract. In a pervasive computing environment, one is facing the problem of handling heterogeneous data from different sources, transmitted over heterogeneous channels and presented on heterogeneous user interfaces. This calls for adaptive data representations keeping as much relevant information as possible while keeping the representation as small as possible. Typically, the gathered data can be high-dimensional vectors with different types of attributes, e.g. continuous, binary and categorical data. In this paper we present- as a first step- a probabilistic latent-variable model, which is capable of fusing high-dimensional heterogenous data into a unified low-dimensional continuous space, and thus brings great benefits for multivariate data analysis, visualization and dimensionality reduction. We adopt a variational approximation to the likelihood of observed data and describe an EM algorithm to fit the model. The advantages of the proposed model are illustrated on toy data and used on real-world painting image data for both visualization and recommendation. 1
Measuring Semantic Relations of Web Sites by Clustering of Local Context.

Dirichlet Enhanced Latent Semantic Analysis.
This paper describes nonparametric Bayesian treatments for analyzing records containing occurrences of items. The introduced model retains the strength of previous approaches that explore the latent factors of each record (e.g. topics of documents), and further uncovers the clustering structure of records, which reflects the statistical dependencies of the latent factors. The nonparametric model induced by a Dirichlet process (DP) flexibly adapts model complexity to reveal the clustering structure of the data. To avoid the problems of dealing with infinite dimensions, we further replace the DP prior by a simpler alternative, namely Dirichlet-multinomial allocation (DMA), which maintains the main modelling properties of the DP. Instead of relying on Markov chain Monte Carlo (MCMC) for inference, this paper applies efficient variational inference based on DMA. The proposed approach yields encouraging empirical results on both a toy problem and text data. The results show that the proposed algorithm uncovers not only the latent factors, but also the clustering structure. 1
Learning Gaussian Process Kernels via Hierarchical Bayes.
We present a novel method for learning with Gaussian process regression in a hierarchical Bayesian framework. In a first step, kernel matrices on a fixed set of input points are learned from data using a simple and efficient EM algorithm. This step is nonparametric, in that it does not require a parametric form of covariance function. In a second step, kernel functions are fitted to approximate the learned covariance matrix using a generalized Nyström method, which results in a complex, data driven kernel. We evaluate our approach as a recommendation engine for art images, where the proposed hierarchical Bayesian method leads to excellent prediction performance. 1
A nonparametric hierarchical bayesian framework for information filtering.
Information filtering has made considerable progress in recent years. The predominant approaches are content-based methods and collaborative methods. Researchers have largely  concentrated on either of the two approaches since a principled unifying framework is still lacking. This paper suggests that both approaches can be combined under a hierarchical Bayesian framework. Individual content-based user profiles are generated and collaboration between various user models is achieved via a common learned prior distribution. However, it turns out that a parametric distribution (e.g. Gaussian) is too restrictive to describe such a common learned prior distribution. We thus introduce a nonparametric common  prior, which is a sample generated from a Dirichlet process which assumes the role of a hyper prior. We describe e#ective means to learn this nonparametric distribution, and apply it to learn users&#039; information needs. The resultant algorithm is simple and understandable, and offers a principled solution to combine content-based filtering and collaborative filtering. Within our framework, we are now able to interpret various existing techniques from a unifying point of view. Finally we demonstrate the empirical success of the proposed information filtering methods.
Probabilistic Memory%2DBased Collaborative Filtering.
Memory-based collaborative filtering (CF) has been studied extensively in the literature and has proven to be successful in various types of personalized recommender systems. In this paper we develop a probabilistic framework for memory-based CF (PMCF). While this framework has clear links with classical memory-based CF, it allows us to find principled solutions to known problems of CF-based recommender systems. In particular, we show that a probabilistic active learning method can be used to actively query the user, thereby solving the &#034;new user problem &#034;. Furthermore, the probabilistic framework allows us to reduce the computational cost of memory-based CF by working on a carefully selected subset of user profiles, while retaining high accuracy. We report experimental results based on two real world data sets, which demonstrate that our proposed PMCF framework allows an accurate and efficient prediction of user preferences.
Knowing a tree from the forest art image retrieval using a society of profiles.
This paper aims to address the problem of art image retrieval (AIR), which aims to help users find their favorite painting images. AIR is of great interests to us because of its application potentials and interesting research challenges--- the retrieval is not only based on painting contents or styles, but also heavily based on user preference profiles. This paper describes the collaborative ensemble learning, a novel statistical learning approach to this task. It at first applies probabilistic support vector machines (SVMs) to model each individual user&#039;s profile based on given examples, i.e. liked or disliked paintings. Due to the high complexity of profile modelling, the SVMs can be rather weak in predicting preferences for new paintings. To overcome this problem, we combine a society of users&#039; profiles, represented by their respective SVM models, to predict a given user&#039;s preferences for painting images. We demonstrate that the combination scheme is embedded in a Bayesian framework and retains intuitive interpretations---like-minded users are likely to share similar preferences. We report extensive empirical studies based on two experimental settings. The first one includes some controlled simulations performed on 4533 painting images. In the second setting, we report evaluations based on user preferences collected through an online web-based survey. Both experiments demonstrate that the proposed approach achieves excellent performance in terms of capturing a user&#039;s diverse preferences.
A Hybrid Relevance%2DFeedback Approach to Text Retrieval.
Relevance feedback (RF) has been an e#ective query modification  approach to improving the performance of information retrieval  (IR) by interactively asking a user whether a set of documents are relevant  or not to a given query concept. The conventional RF algorithms  either converge slowly or cost a user&#039;s additional e#orts in reading irrelevant  documents. This paper surveys several RF algorithms and introduces  a novel hybrid RF approach using a support vector machine  (HRFSVM), which actively selects the uncertain documents as well as  the most relevant ones on which to ask users for feedback. It can efficiently  rank documents in a natural way for user browsing. We conduct  experiments on Reuters-21578 dataset and track the precision as  a function of feedback iterations. Experimental results have shown that  HRFSVM significantly outperforms two other RF algorithms.
Representative Sampling for Text Classification Using Support Vector Machines.
Abstract. In order to reduce human efforts, there has been increasing interest in applying active learning for training text classifiers. This paper describes a straightforward active learning heuristic, representative sampling, which explores the clustering structure of ‘uncertain ’ documents and identifies the representative samples to query the user opinions, for the purpose of speeding up the convergence of Support Vector Machine (SVM) classifiers. Compared with other active learning algorithms, the proposed representative sampling explicitly addresses the problem of selecting more than one unlabeled documents. In an empirical study we compared representative sampling both with random sampling and with SVM active learning. The results demonstrated that representative sampling offers excellent learning performance with fewer labeled documents and thus can reduce human efforts in text classification tasks. 1
An Introduction to Nonparametric Hierarchical Bayesian Modelling with a Focus on Multi%2Dagent Learning.

Collaborative Ensemble Learning Combining Collaborative and Content%2DBased Information Filtering via Hierarchical Bayes.
Collaborative filtering (CF) and contentbased filtering (CBF) have widely been used in information filtering applications, both approaches having their individual strengths and weaknesses. This paper proposes a novel probabilistic framework to unify CF and CBF, named collaborative ensemble learning. Based on content based probabilistic models for each user’s preferences (the CBF idea), it combines a society of users ’ preferences to predict an active user’s preferences (the CF idea). While retaining an intuitive explanation, the combination scheme can be interpreted as a hierarchical Bayesian approach in which a common prior distribution is learned from related experiments. It does not require a global training stage and thus can incrementally incorporate new data. We report results based on two data sets, the Reuters-21578 text data set and a data base of user opionions on art images. For both data sets, collaborative ensemble achieved excellent performance in terms of recommendation accuracy. In addition to recommendation engines, collaborative ensemble learning is applicable to problems typically solved via classical hierarchical Bayes, like multisensor fusion and multitask learning. 1

Feature Weighting and Instance Selection for Collaborative Filtering An Information%2DTheoretic Approach%2A.

Removing redundancy and inconsistency in memory%2Dbased collaborative filtering.
The application range of memory-based collaborative filtering (CF) is limited due to CF&#039;s high memory consumption and long runtime. The approach presented in this paper removes redundant and inconsistent instances (users) from the data. Our work shows that a satisfactory accuracy can be achieved by using only a small portion of the original data set, thereby alleviating the storage and runtime cost of the CF algorithm. In our approach, we consider instance selection as the problem of selecting informative data that increase the a posteriori probability of the optimal model. We evaluate the empirical performance of our approach on two realworld data sets and attain very promising results. Data size and prediction time are significantly reduced, while the prediction accuracy is on a par with results achieved by using the complete database.
Instance Selection Techniques for Memory%2Dbased Collaborative Filtering.
Abstract: Collaborative filtering (CF) has become an important data mining technique to make personalized recommendations for books, web pages or movies, etc. One popular algorithm is the memory-based collaborative filtering, which predicts a user’s preference based on his or her similarity to other users (instances) in the database. However, the tremendous growth of users and the large number of products, memory-based CF algorithms results in the problem of deciding the right instances to use during prediction, in order to reduce executive cost and excessive storage, and possibly to improve the generalization accuracy by avoiding noise and overfitting. In this paper, we focus our work on a typical user preference database that contains many missing values, and propose four novel instance reduction techniques called TURF1-TURF4 as a preprocessing step to improve the efficiency and accuracy of the memory-based CF algorithm. The key idea is to generate prediction from a carefully selected set of relevant instances. We evaluate the techniques on the well-known EachMovie data set. Our experiments showed that the proposed algorithms not just dramatically speed up the prediction, but also improved the accuracy. 1

Models for MIMO propagation channels a review.
This paper reviews recently published results on multiple input multiple output (MIMO) channel modeling. Both narrowband and wideband models are considered. We distinguish between two main approaches to MIMO channel modeling, namely, physically based and non-physically based modeling. The non-physical models primarily rely on the statistical characteristics of the MIMO channels obtained from the measured data while the physical models describe the MIMO channel (or its distribution) via some physical parameters. We briefly review different MIMO channel models and discuss their relationships. Some interesting aspects will be described in more detail and we note areas where few results are available
Selecting Relevant Instances for Efficient and Accurate Collaborative Filtering.
Collaborative filtering uses a database about consumers&#039; preferences to make personal product recommendations and is achieving widespread success in both E-Commerce and Information Filtering Applications nowadays. However, the traditional collaborative filtering algorithms do not scale well to the ever-growing number of consumers. The quality of the recommendation also needs to be improved in order to gain more trust from the consumers. In this paper, we present a novel method to improve the scalability and the accuracy of the collaborative filtering algorithm. We introduce an information theoretic approach to measure the relevance of a consumer (instance) for predicting the preference for the given product (target concept). The proposed method reduces the training data set by selecting only highly relevant instances. Our experimental evaluation on the well-known EachMovie data set shows that our method doesn&#039;t only significantly speed up the prediction, but also results in a better accuracy.
Feature Weighting Methods for Improving the Accuracy of Collaborative Filtering %28Invited Address%29.

