Mixed group ranks: preference and confidence in classifier combination

Classifier combination holds the potential of improving performance by combining the results of multiple classifiers. For domains with very large numbers of classes, such as biometrics, we present an axiomatic framework of desirable mathematical properties for combination functions of rank-based classifiers. This framework represents a continuum of combination rules, including the Borda Count, Logistic Regression, and Highest Rank combination methods as extreme cases. Intuitively, this framework captures how the two complementary concepts of general preference for specific classifiers and the confidence it has in any specific result (as indicated by ranks) can be balanced while maintaining consistent rank interpretation. Mixed Group Ranks (MGR) is a new combination function that balances preference and confidence by generalizing these other functions. We demonstrate that MGR is an effective combination approach by performing multiple experiments on data sets with large numbers of classes and classifiers from the FERET face recognition study.     

赛灵思节一代无线技术创新

文章介绍了赛灵思公司的LTE基站目标设计平台、无线基础架构的半导体解决方案,以及用于射频拉远和射频卡的LTE数字前端的LTE解决方案.     

Using Lyapunov's direct method for wave suppression in reactive systems

This paper proposes a new approach for stabilizing a homogeneous solution in reaction–convection–diffusion system with oscillatory kinetics, in which moving or stationary patterns emerge in the absence of control. Specifically, we aim to suppress patterns by using a spatially weighted finite-dimensional feedback control that assures stability of the solution according to Lyapunov's direct method. A practical design procedure, based on spectral representation of the system and dissipative nature of parabolic PDEs, is presented.     

TEG—a hybrid approach to information extraction

This paper describes a hybrid statistical and knowledge-based information extraction model, able to extract entities and relations at the sentence level. The model attempts to retain and improve the high accuracy levels of knowledge-based systems while drastically reducing the amount of manual labour by relying on statistics drawn from a training corpus. The implementation of the model, called TEG (trainable extraction grammar), can be adapted to any IE domain by writing a suitable set of rules in a SCFG (stochastic context-free grammar)-based extraction language and training them using an annotated corpus. The system does not contain any purely linguistic components, such as PoS tagger or shallow parser, but allows to using external linguistic components if necessary. We demonstrate the performance of the system on several named entity extraction and relation extraction tasks. The experiments show that our hybrid approach outperforms both purely statistical and purely knowledge-based systems, while requiring orders of magnitude less manual rule writing and smaller amounts of training data. We also demonstrate the robustness of our system under conditions of poor training-data quality. Ronen Feldman is a senior lecturer at the Mathematics and Computer Science Department of Bar-Ilan University in Israel, and the Director of the Data Mining Laboratory. He received his B.Sc. in Math, Physics and Computer Science from the Hebrew University, M.Sc. in Computer Science from Bar-Ilan University, and his Ph.D. in Computer Science from Cornell University in NY. He was an Adjunct Professor at NYU Stern Business School. He is the founder of ClearForest Corporation, a Boston based company specializing in development of text mining tools and applications. He has given more than 30 tutorials on next mining and information extraction and authored numerous papers on these topics. He is currently finishing his book “The Text Mining Handbook” to the published by Cambridge University Press. Benjamin Rosenfeld is a research scientist at ClearForest Corporation. He received his B.Sc. in Mathematics and Computer Science from Bar-Ilan University. He is the co-inventor of the DIAL information extraction language. Moshe Fresko is finalizing his Ph.D. in Computer Science Department at Bar-Ilan University in Israel. He received his B.Sc. in Computer Engineering from Bogazici University, Istanbul/Turkey on 1991, and M.Sc. on 1994. He is also an adjunct lecturer at the Computer Science Department of Bar-Ilan University and functions as the Information-Extraction Group Leader in the Data Mining Laboratory.     

Optimized temporal monitors for SystemC

SystemC is a modeling language built as an extension of C++. Its growing popularity and the increasing complexity of designs have motivated research efforts aimed at the verification of SystemC models using assertion-based verification (ABV), where the designer asserts properties that capture the design intent in a formal language such as PSL or SVA. The model then can be verified against the properties using runtime or formal verification techniques. In this paper we focus on automated generation of runtime monitors from temporal properties. Our focus is on minimizing runtime overhead, rather than monitor size or monitor-generation time. We identify four issues in monitor generation: state minimization, alphabet representation, alphabet minimization, and monitor encoding. We conduct extensive experimentation and identify a combination of settings that offers the best performance in terms of runtime overhead.     

2010 CAV award announcement

The 2010 CAV (Computer-Aided Verification) award was awarded to Kenneth L. McMillan of Cadence Research Laboratories for a series of fundamental contributions resulting in significant advances in scalability of model checking tools. The annual award recognizes a specific fundamental contribution or a series of outstanding contributions to the CAV field.     

Synthesis from component libraries

Synthesis is the automated construction of a system from its specification. In the classical temporal synthesis algorithms, it is always assumed the system is “constructed from scratch” rather than “composed” from reusable components. This, of course, rarely happens in real life. In real life, almost every non-trivial commercial system, either in hardware or in software system, relies heavily on using libraries of reusable components. Furthermore, other contexts, such as web-service orchestration, can be modeled as synthesis of a system from a library of components. In this work, we define and study the problem of LTL synthesis from libraries of reusable components. We define two notions of composition: data-flow composition, for which we prove the problem is undecidable, and control-flow composition, for which we prove the problem is 2EXPTIME-complete. As a side benefit, we derive an explicit characterization of the information needed by the synthesizer on the underlying components. This characterization can be used as a specification formalism between component providers and integrators.     

Combinatorial Optimization and Performance Analysis of a Multi-arm Cartesian Robotic Fruit Harvester—Extensions of Graph Coloring

A mobile melon robotic harvester consisting of multiple Cartesian manipulators, each with three degrees of freedom, is being developed. In order to design an optimal robot in terms of number of arms, manipulator capabilities, and robot speed, a method of allocating the fruits to be picked by each manipulator in a way that yields the maximum harvest has been developed. Such a method has already been devised for a multi-arm robot with 2DOF each. The maximum robotic harvesting problem was shown there to be an example of the maximum k-colorable subgraph problem (MKCSP) on an interval graph. However, for manipulators with 3DOF, the additional longitudinal motion results in variable intervals. To overcome this issue, we devise a new model based on the color-dependent interval graph (CDIG). This enables the harvest by multiple robotic arms to be modeled as a modified version of the MKCSP. Based on previous research, we develop a greedy algorithm that solves the problem in polynomial time, and prove its optimality using induction. As with the multi-arm 2DOF robot, when simulated numerous times on a field of randomly distributed fruits, the algorithm yields a nearly identical percentage of fruit harvested for given robot parameters. The results of the probabilistic analysis developed for the 2DOF robot was modified to yield a formula for the expected harvest ratio of the 3DOF robot. The significance of this method is that it enables selecting the most efficient actuators, number of manipulators, and robot forward velocity for maximal robotic fruit harvest.     

An exact high order beam element

This work gives the exact stiffness coefficients for an high order isotropic beam element. The terms are found directly from the solutions of the differential equations that describe the deformations of the cross-section according to the high order theory, which include cubic variation of the axial displacements over the cross-section of the beam. The model has six degrees of freedom at the two ends, one transverse displacement and two rotations, and the end forces are a shear force and two end moments. Also given are the equivalent end forces and moments for several cases of loading along the member. The components of the end moments are investigated, and are found for exact results. Comparison is made with the Bernoulli-Euler and Timoshenko beam models.     

The effects of divided attention on encoding processes in memory: Mapping the locus of interference.

Despite the detrimental effects of divided attention at encoding on later memory performance, results described in the literature do not unequivocally specify which processes are interrupted during encoding by participants' occupation with a concurrent task. Using a processing analysis framework where the encoding process is viewed as a multiphase mental activity, the current research investigated this issue using a new differential temporal interference paradigm where the study phase of single words was interrupted at different temporal segments. In two experiments, we used performance on both memory and online choice reaction time tasks to assess whether such differential interference would produce different degrees of reduction in participants' later memory performance, as well as changes in the attentional resources required to execute each of the encoding phases. Measures of memory and concurrent task performance in the two experiments converged on similar patterns, showing that all phases of encoding are affected by the concurrent task. However, the initial encoding phase, which is tentatively associated with the initial registration of information, seems especially vulnerable to interference. (PsycINFO Database Record (c) 2010 APA, all rights reserved)