We present a novel model-metric co-learning (MMCL) methodology for sequence classification which learns in the model space -- each data item (sequence) is represented by a predictive model from a carefully designed model class. MMCL learning encourages sequences from the same class to be represented by ‘close’ model representations, well separated from those for different classes. Existing approaches to the problem either fit a single model to all the data, or a (predominantly linear) model on each sequence. We introduce a novel hybrid approach spanning the two extremes. The model class we use is a special form of adaptive high-dimensional non-linear state space model with a highly constrained and simple dynamic part. The dynamic part is identical for all data items and acts as a temporal filter providing a rich pool of dynamic features that can be selectively extracted by individual (static) linear readout mappings representing the sequences. Alongside learning the dynamic part, we also learn the global metric in the model readout space. Experiments on synthetic and benchmark data sets confirm the effectiveness of the algorithm compared to a variety of alternative methods.

The k-means algorithm is one of the most often used method for data clustering. However, the standard k-means can only be applied in the original feature space. The kernel k-means, which extends k-means into the kernel space, can be used to capture the non-linear structure and identify arbitrarily shaped clusters. Since both the standard k-means and kernel k-means apply the squared error to measure the distances between data points and cluster centers, a few outliers will cause large errors and dominate the objection function. Besides, the performance of kernel method is largely determined by the choice of kernel. Unfortunately, the most suitable kernel for a particular task is often unknown in advance. In this paper, we first present a robust k-means using l 2,1 -norm in the feature space and then extend it to the kernel space. To recap the powerfulness of kernel methods, we further propose a novel robust multiple kernel k-means (RMKKM) algorithm that simultaneously finds the best clustering label, the cluster membership and the optimal combination of multiple kernels. An alternating iterative schema is developed to find the optimal value. Extensive experiments well demonstrate the effectiveness of the proposed algorithms.

Learning from crowds , where the labels of data instances are collected using a crowdsourcing way, has attracted much attention during the past few years. In contrast to a typical crowdsourcing setting where all data instances are assigned to annotators for labeling,  active learning from crowds actively selects a subset of data instances and assigns them to the annotators, thereby reducing the cost of labeling. This paper goes a step further. Rather than assume all annotators must provide labels, we allow the annotators to express that they are unsure about the assigned data instances. By adding the “unsure” option, the workloads for the annotators are somewhat reduced, because saying “unsure” will be easier than trying to provide a crisp label for some difficult data instances. Moreover, it is safer to use “unsure” feedback than to use labels from reluctant annotators because the latter has more chance to be misleading. Furthermore, different annotators may experience difficulty in different data instances, and thus the unsure option provides a valuable ingredient for modeling crowds’ expertise. We propose the ALCU-SVM algorithm for this new learning problem. Experimental studies on simulated and real crowdsourcing data show that, by exploiting the unsure option, ALCU-SVM achieves very promising performance.

Incorporating multiple types of relational information from heterogeneous networks has been proved effective in data mining. Although Wikipedia is one of the most famous heterogeneous network, previous works of semantic analysis on Wikipedia are mostly limited on single type of relations. In this paper, we aim at incorporating multiple types of relations to measure the semantic relatedness between Wikipedia entities. We propose a framework of coordinate matrix factorization to construct low-dimensional continuous representation for entities, categories and words in the same semantic space. We formulate this task as the completion of a sparse entity-entity association matrix, in which each entry quantifies the strength of relatedness between corresponding entities. We evaluate our model on the task of judging pair-wise word similarity. Experiment result shows that our model outperforms both traditional entity relatedness algorithms and other representation learning models.

Representation learning has shown its effectiveness in many tasks such as image classification and text mining. Network representation learning aims at learning distributed vector representation for each vertex in a network, which is also increasingly recognized as an important aspect for network analysis. Most network representation learning methods investigate network structures for learning. In reality, network vertices contain rich information (such as text), which cannot be well applied with algorithmic frameworks of typical representation learning methods. By proving that DeepWalk, a state-of-the-art network representation method, is actually equivalent to matrix factorization (MF), we propose text-associated DeepWalk (TADW). TADW incorporates text features of vertices into network representation learning under the framework of matrix factorization. We evaluate our method and various baseline methods by applying them to the task of multi-class classification of vertices. The experimental results show that, our method outperforms other baselines on all three datasets, especially when networks are noisy and training ratio is small.

Retrieving similar questions is very important in community-based question answering. A major challenge is the lexical gap in sentence matching. In this paper, we propose a convolutional neural tensor network architecture to encode the sentences in semantic space and model their interactions with a tensor layer. Our model integrates sentence modeling and semantic matching into a single model, which can not only capture the useful information with convolutional and pooling layers, but also learn the matching metrics between the question and its answer. Besides, our model is a general architecture, with no need for the other knowledge such as lexical or syntactic analysis. The experimental results shows that our method outperforms the other methods on two matching tasks.

In this paper, we investigate the problem of face clustering in real-world videos. In many cases, the distribution of the face data is unbalanced. In movies or TV series videos, the leading casts appear quite often and the others appear much less. However, many clustering algorithms cannot well handle such severe unbalance between the data distribution, resulting in that the large class is split apart, and the small class is merged into the large ones and thus missing. On the other hand, the data distribution proportion information may be known beforehand. For example, we can obtain such information by counting the spoken lines of the characters in the script text. Hence, we propose to make use of the proportion prior to regularize the clustering. A Hidden Conditional Random Field(HCRF) model is presented to incorporate the proportion prior. In experiments on a public data set from real-world videos, we observe improvements on clustering performance against state-of-the-art methods.

Differential privacy preserving regression models guarantee protection against attempts to infer whether a subject was included in the training set used to derive a model. It is not designed to protect attribute privacy of a target individual when model inversion attacks are launched. In model inversion attacks, an adversary uses the released model to make predictions of sensitive attributes (used as input to the model) of a target individual when some background information about the target individual is available. Previous research showed that existing differential privacy mechanisms cannot effectively prevent model inversion attacks while retaining model efficacy. In this paper, we develop a novel approach which leverages the functional mechanism to perturb coefficients of the polynomial representation of the objective function but effectively balances the privacy budget for sensitive and non-sensitive attributes in learning the differential privacy preserving regression model. Theoretical analysis and empirical evaluations demonstrate our approach can effectively prevent model inversion attacks and retain model utility.

Argumentation mining aims to automatically identify structured argument data from unstructured natural language text. This challenging, multi-faceted task is recently gaining a growing attention, especially due to its many potential applications. One particularly important aspect of argumentation mining is claim identification. Most of the current approaches are engineered to address specific domains. However, argumentative sentences are often characterized by common rhetorical structures, independently of the domain. We thus propose a method that exploits structured parsing information to detect claims without resorting to contextual information, and yet achieve a performance comparable to that of state-of-the-art methods that heavily rely on the context.

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.