A Simple Label Switching Algorithm for Semisupervised Structural SVMs

In structured output learning, obtaining labeled data for real-world applications is usually costly, while unlabeled examples are available in abundance. Semisupervised structured classification deals with a small number of labeled examples and a large number of unlabeled structured data. In this work, we consider semisupervised structural support vector machines with domain constraints. The optimization problem, which in general is not convex, contains the loss terms associated with the labeled and unlabeled examples, along with the domain constraints. We propose a simple optimization approach that alternates between solving a supervised learning problem and a constraint matching problem. Solving the constraint matching problem is difficult for structured prediction, and we propose an efficient and effective label switching method to solve it. The alternating optimization is carried out within a deterministic annealing framework, which helps in effective constraint matching and avoiding poor local minima, which are not very useful. The algorithm is simple and easy to implement. Further, it is suitable for any structured output learning problem where exact inference is available. Experiments on benchmark sequence labeling data sets and a natural language parsing data set show that the proposed approach, though simple, achieves comparable generalization performance.

New phase unwrapping strategy for rapid and dense 3D data acquisition in structured light approach

Sinusoidal structured light projection (SSLP) technique, specifically-phase stepping method, is in widespread use to obtain accurate, dense 3-D data. But, if the object under investigation possesses surface discontinuities, phase unwrapping (an intermediate step in SSLP) stage mandatorily require several additional images, of the object with projected fringes (of different spatial frequencies), as input to generate a reliable 3D shape. On the other hand, Color-coded structured light projection (CSLP) technique is known to require a single image as in put, but generates sparse 3D data. Thus we propose the use of CSLP in conjunction with SSLP to obtain dense 3D data with minimum number of images as input. This approach is shown to be significantly faster and reliable than temporal phase unwrapping procedure that uses a complete exponential sequence. For example, if a measurement with the accuracy obtained by interrogating the object with 32 fringes in the projected pattern is carried out with both the methods, new strategy proposed requires only 5 frames as compared to 24 frames required by the later method.

DFL: A Data Flow Language

Many novel computer architectures like array and multiprocessors which achieve high performance through the use of concurrency exploit variations of the von Neumann model of computation. The effective utilization of the machines makes special demands on programmers and their programming languages, such as the structuring of data into vectors or the partitioning of programs into concurrent processes. In comparison, the data flow model of computation demands only that the principle of structured programming be followed. A data flow program, often represented as a data flow graph, is a program that expresses a computation by indicating the data dependencies among operators. A data flow computer is a machine designed to take advantage of concurrency in data flow graphs by executing data independent operations in parallel. In this paper, we discuss the design of a high level language (DFL: Data Flow Language) suitable for data flow computers. Some sample procedures in DFL are presented. The implementation aspects have not been discussed in detail since there are no new problems encountered. The language DFL embodies the concepts of functional programming, but in appearance closely resembles Pascal. The language is a better vehicle than the data flow graph for expressing a parallel algorithm. The compiler has been implemented on a DEC 1090 system in Pascal.

Dielectric relaxation in bismuth layer-structured BaBi4Ti4O15 ferroelectric ceramics

The dielectric properties of BaBi4Ti4O15 ceramics were investigated as a function of frequency (10(2)-10(6) Hz) at various temperatures (30 degrees C-470 degrees C), covering the phase transition temperature. Two different conduction mechanisms were obtained by fitting the complex impedance data to Cole-Cole equation. The grain and grain boundary resistivities were found to follow the Arrhenius law associated with activation energies: E-g similar to 1.12 eV below T-m and E-g similar to 0.70 eV above T-m for the grain conduction; and E-gb similar to 0.93 eV below T-m and E-gb similar to 0.71 eV above T-m for the grain boundary conduction. Relaxation times extracted using imaginary part of complex impedance Z `'(omega) and modulus M `'(omega) were also found to follow the Arrhenius law and showed an anomaly around the phase transition temperature. The frequency dependence of conductivity was interpreted in terms of the jump relaxation model and was fitted to the double power law. (C) 2010 Elsevier B. V. All rights reserved.

Efficient statistical modeling for the compression of tree structured intermediate code

We propose a scheme for the compression of tree structured intermediate code consisting of a sequence of trees specified by a regular tree grammar. The scheme is based on arithmetic coding, and the model that works in conjunction with the coder is automatically generated from the syntactical specification of the tree language. Experiments on data sets consisting of intermediate code trees yield compression ratios ranging from 2.5 to 8, for file sizes ranging from 167 bytes to 1 megabyte.

Structured Learning for Non-Smooth Ranking losses

Learning to rank from relevance judgment is an active research area. Itemwise score regression, pairwise preference satisfaction, and listwise structured learning are the major techniques in use. Listwise structured learning has been applied recently to optimize important non-decomposable ranking criteria like AUC (area under ROC curve) and MAP(mean average precision). We propose new, almost-lineartime algorithms to optimize for two other criteria widely used to evaluate search systems: MRR (mean reciprocal rank) and NDCG (normalized discounted cumulative gain)in the max-margin structured learning framework. We also demonstrate that, for different ranking criteria, one may need to use different feature maps. Search applications should not be optimized in favor of a single criterion, because they need to cater to a variety of queries. E.g., MRR is best for navigational queries, while NDCG is best for informational queries. A key contribution of this paper is to fold multiple ranking loss functions into a multi-criteria max-margin optimization.The result is a single, robust ranking model that is close to the best accuracy of learners trained on individual criteria. In fact, experiments over the popular LETOR and TREC data sets show that, contrary to conventional wisdom, a test criterion is often not best served by training with the same individual criterion.

Scalable multi-dimensional user intent identification using tree structured distributions

The problem of identifying user intent has received considerable attention in recent years, particularly in the context of improving the search experience via query contextualization. Intent can be characterized by multiple dimensions, which are often not observed from query words alone. Accurate identification of Intent from query words remains a challenging problem primarily because it is extremely difficult to discover these dimensions. The problem is often significantly compounded due to lack of representative training sample. We present a generic, extensible framework for learning the multi-dimensional representation of user intent from the query words. The approach models the latent relationships between facets using tree structured distribution which leads to an efficient and convergent algorithm, FastQ, for identifying the multi-faceted intent of users based on just the query words. We also incorporated WordNet to extend the system capabilities to queries which contain words that do not appear in the training data. Empirical results show that FastQ yields accurate identification of intent when compared to a gold standard.

Adaptive executions of hyperbolic block-structured AMR applications on GPU systems

A block-structured adaptive mesh refinement (AMR) technique has been used to obtain numerical solutions for many scientific applications. Some block-structured AMR approaches have focused on forming patches of non-uniform sizes where the size of a patch can be tuned to the geometry of a region of interest. In this paper, we develop strategies for adaptive execution of block-structured AMR applications on GPUs, for hyperbolic directionally split solvers. While effective hybrid execution strategies exist for applications with uniform patches, our work considers efficient execution of non-uniform patches with different workloads. Our techniques include bin-packing work units to load balance GPU computations, adaptive asynchronism between CPU and GPU executions using a knapsack formulation, and scheduling communications for multi-GPU executions. Our experiments with synthetic and real data, for single-GPU and multi-GPU executions, on Tesla S1070 and Fermi C2070 clusters, show that our strategies result in up to a 3.23 speedup in performance over existing strategies.