Part deformation has been a longstanding challenge for object parsing, of which the primary difficulty lies in modeling the highly diverse object structures. To this end, we propose a novel structure parsing model to capture deformable object structures. The proposed model consists of two de-formable layers: the top layer is an undirected graph that incorporates inter-part deformations to infer object structures; the base layer is consisted of various independent nodes to characterize local intra-part deformations. To learn this two-layer model, we design a layer-wise learning algorithm,which employs matching pursuit and belief propagation for a low computational complexity inference. Specifically, active basis sparse coding is leveraged to build the nodes at the base layer, while the edge weights are estimated by a structural support vector machine. Experimental results on two benchmark datasets (i.e., faces and horses) demonstrate that the proposed model yields superior parsing performance over state-of-the-art models.

In this paper, we propose a new local spatio-temporal descriptor for videos and we propose a new approach for action recognition in videos based on the introduced descriptor. The new descriptor is called the Video Covariance Matrix Logarithm (VCML). The VCML descriptor is based on a covariance matrix representation, and it models relationships between different low-level features, such as intensity and gradient. We apply the VCML descriptor to encode appearance information of local spatio-temporal video volumes, which are extracted by the Dense Trajectories. Then, we present an extensive evaluation of the proposed VCML descriptor with the Fisher vector encoding and the Support Vector Machines on four challenging action recognition datasets. We show that the VCML descriptor achieves better results than the state-of-the-art appearance descriptors. Moreover, we present that the VCML descriptor carries complementary information to the HOG descriptor and their fusion gives a significant improvement in action recognition accuracy. Finally, we show that the VCML descriptor improves action recognition accuracy in comparison to the state-of-the-art Dense Trajectories, and that the proposed approach achieves superior performance to the state-of-the-art methods.

Continuous optimization is an important problem in many areas of AI, including vision, robotics, probabilistic inference, and machine learning. Unfortunately, most real-world optimization problems are nonconvex, causing standard convex techniques to find only local optima, even with extensions like random restarts and simulated annealing. We observe that, in many cases, the local modes of the objective function have combinatorial structure, and thus ideas from combinatorial optimization can be brought to bear. Based on this, we propose a problem-decomposition approach to nonconvex optimization. Similarly to DPLL-style SAT solvers and recursive conditioning in probabilistic inference, our algorithm, RDIS, recursively sets variables so as to simplify and decompose the objective function into approximately independent sub-functions, until the remaining functions are simple enough to be optimized by standard techniques like gradient descent. The variables to set are chosen by graph partitioning, ensuring decomposition whenever possible. We show analytically that RDIS can solve a broad class of nonconvex optimization problems exponentially faster than gradient descent with random restarts. Experimentally, RDIS outperforms standard techniques on problems like structure from motion and protein folding.

The travelling thief problem (TTP) is a multi-component optimisation problem involving two interdependent NP-hard components: the travelling salesman problem (TSP) and the knapsack problem (KP). Recent state-of-the-art TTP solvers modify the underlying TSP and KP solutions in an iterative and interleaved fashion. The TSP solution (cyclic tour) is typically changed in a deterministic way, while changes to the KP solution typically involve a random search, effectively resulting in a quasi-meandering exploration of the TTP solution space. Once a plateau is reached, the iterative search of the TTP solution space is restarted by using a new initial TSP tour. We propose to make the search more efficient though an adaptive surrogate model (based on a customised form of Support Vector Regression) that learns the characteristics of initial TSP tours that lead to good TTP solutions.

In this paper, we aim to predict students' learning perfor-mance by combining two-modality sensing variables, namely eye tracking that monitors learners' eye movements and elec-troencephalography (EEG) that measures learners' cerebral activity. Our long-term goal is to use both data to provide ap-propriate adaptive assistance for students to enhance their learning experience and optimize their performance. An ex-perimental study was conducted in order to collet gaze data and brainwave signals of fifteen students during an interac-tion with a virtual learning environment. Different classifica-tion algorithms were used to discriminate between two groups of learners: students who successfully resolve the problem-solving tasks and students who do not. Experimental results demonstrated that the K-Nearest Neighbor classifier achieved good accuracy when combining both eye movement and EEG features compared to using solely eye movement or EEG.

In this work, we first attempt to replicate an earlier study on gene selection and clustering, and then we extend this work by applying a different type of hierarchical clustering to dis- cover interesting subsets of genes from breast cancer data. Replication of such studies is a known challenge and an ac- tive area of research in bioinformatics. The work presented in this paper is three-fold. First, we replicate a study conducted at the University of North Carolina to generate an initial set of genes. Second, we apply an approach called Distance Weighted Discrimination to fuse multiple, disparate breast cancer datasets into a single validation set. Third, we per- form hierarchical clustering and k-means clustering on this validation set to discover natural groupings and compare the clusters generated by both methods. While applying the hi- erarchical clustering is part of the reproduction step, we ex- tend the research by trying two different forms of hierarchi- cal clustering. We also apply k-means clustering for the same purpose and compare all three methods using Kaplan-Meier estimation and Cox proportional hazards regression. We dis- cover that among the three methods, k-means clustering gives us the best results.

Dichotomy transformation in biometric authentication problem creates a two class (""within"" or ""between"") classification problem in multivariate distance space. Linear discriminant analysis, which is a linear classifier, results in good performance in IRIS biometric authentication problem. However, it assumes that the distributions of two classes are normal, whereas they are closely related to the log-normal distributions. Here a modified variance linear discriminant analysis algorithm is proposed and its superior experimental results on the IRIS biometric database are reported.

We set out to compare the utility of different representations of driving simulator time series data in the context of both supervised and unsupervised learning algorithms. Given the task of identifying similar time series; it is important to understand how a dataset of time series samples might be distributed and how effectively different methods capture the groupings of distinct behaviors. First we engineer three representations of the driving simulator data: converting them to feature vectors, using the raw time series, and rendering them as images. At which point, we introduce a novel method for comparing time series using temporal and spatial alignments. Then, we employ a battery of clustering algorithms to isolate groups of samples with similar traits and evaluate the quality of clusters produced.

The objective of my doctoral research is bring together two fields: partially-observable reinforcement learning (PORL) and non-parametric Bayesian statistics (NPB) to address issues of statistical modeling and decision-making in complex, real-world domains.

Unhealthy diet plays a major role in driving chronic disease incidence and prevalence. Taxation of unhealthy food has been proposed to improve population-level dietary patterns, and its effectiveness can be estimated by the prediction of the change in unhealthy food purchasing upon increase of food price. Recent availability of grocery transaction data from scanner technologies enables an accurate prediction of food sales. However, the very large number of product at-tributes in these data prohibits the application of conventional statistical learning algorithms. In this study, we explored the predictive performance of learning algorithms adapted for high-dimensional data, namely the Least Absolute Shrinkage and Selection Operator (LASSO) and Decision Tree Regressor with Adaptive Boosting (DTR-Ada-Boost), in comparison with a conventional statistical learning based on Ordinary Least Square (OLS). LASSO demonstrated superior predictive accuracy to OLS, possibly due to its ability to reduce over fitting and collinearity across predictive features of food sales. DTR-AdaBoost showed the best predictive accuracy, suggesting the presence of extensive non-linearity between the predictive features in the transaction data and sales.