Results 1  10
of
12
Fast SVM training algorithm with decomposition on very large data sets
 IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
, 2005
"... Training a support vector machine on a data set of huge size with thousands of classes is a challenging problem. This paper proposes an efficient algorithm to solve this problem. The key idea is to introduce a parallel optimization step to quickly remove most of the nonsupport vectors, where block ..."
Abstract

Cited by 20 (2 self)
 Add to MetaCart
Training a support vector machine on a data set of huge size with thousands of classes is a challenging problem. This paper proposes an efficient algorithm to solve this problem. The key idea is to introduce a parallel optimization step to quickly remove most of the nonsupport vectors, where block diagonal matrices are used to approximate the original kernel matrix so that the original problem can be split into hundreds of subproblems which can be solved more efficiently. In addition, some effective strategies such as kernel caching and efficient computation of kernel matrix are integrated to speed up the training process. Our analysis of the proposed algorithm shows that its time complexity grows linearly with the number of classes and size of the data set. In the experiments, many appealing properties of the proposed algorithm have been investigated and the results show that the proposed algorithm has a much better scaling capability than Libsvm, SVM light, and SVMTorch. Moreover, the good generalization performances on several large databases have also been achieved.
Grounding Analog Computers
 Think
, 1993
"... Although analog computation was eclipsed by digital computation in the second half of the twentieth century, it is returning as an important alternative computing technology. Indeed, as explained in this report, theoretical results imply that analog computation can escape from the limitations of dig ..."
Abstract

Cited by 12 (7 self)
 Add to MetaCart
Although analog computation was eclipsed by digital computation in the second half of the twentieth century, it is returning as an important alternative computing technology. Indeed, as explained in this report, theoretical results imply that analog computation can escape from the limitations of digital computation. Furthermore, analog computation has emerged as an important theoretical framework for discussing computation in the brain and other natural systems. The report (1) summarizes the fundamentals of analog computing, starting with the continuous state space and the various processes by which analog computation can be organized in time; (2) discusses analog computation in nature, which provides models and inspiration for many contemporary uses of analog computation, such as neural networks; (3) considers generalpurpose analog computing, both from a theoretical perspective and in terms of practical generalpurpose analog computers; (4) discusses the theoretical power of
A survey on continuous time computations
 New Computational Paradigms
"... Abstract. We provide an overview of theories of continuous time computation. These theories allow us to understand both the hardness of questions related to continuous time dynamical systems and the computational power of continuous time analog models. We survey the existing models, summarizing resu ..."
Abstract

Cited by 11 (2 self)
 Add to MetaCart
Abstract. We provide an overview of theories of continuous time computation. These theories allow us to understand both the hardness of questions related to continuous time dynamical systems and the computational power of continuous time analog models. We survey the existing models, summarizing results, and point to relevant references in the literature. 1
Upper and Lower Bounds on ContinuousTime Computation
"... We consider various extensions and modifications of Shannon's General Purpose Analog Computer, which is a model of computation by differential equations in continuous time. We show that several classical computation classes have natural analog counterparts, including the primitive recursive function ..."
Abstract

Cited by 8 (2 self)
 Add to MetaCart
We consider various extensions and modifications of Shannon's General Purpose Analog Computer, which is a model of computation by differential equations in continuous time. We show that several classical computation classes have natural analog counterparts, including the primitive recursive functions, the elementary functions, the levels of the Grzegorczyk hierarchy, and the arithmetical and analytical hierarchies.
Decidability and universality in symbolic dynamical systems
 Fund. Inform
"... Abstract. Many different definitions of computational universality for various types of dynamical systems have flourished since Turing’s work. We propose a general definition of universality that applies to arbitrary discrete time symbolic dynamical systems. Universality of a system is defined as un ..."
Abstract

Cited by 5 (0 self)
 Add to MetaCart
Abstract. Many different definitions of computational universality for various types of dynamical systems have flourished since Turing’s work. We propose a general definition of universality that applies to arbitrary discrete time symbolic dynamical systems. Universality of a system is defined as undecidability of a modelchecking problem. For Turing machines, counter machines and tag systems, our definition coincides with the classical one. It yields, however, a new definition for cellular automata and subshifts. Our definition is robust with respect to initial condition, which is a desirable feature for physical realizability. We derive necessary conditions for undecidability and universality. For instance, a universal system must have a sensitive point and a proper subsystem. We conjecture that universal systems have infinite number of subsystems. We also discuss the thesis according to which computation should occur at the ‘edge of chaos ’ and we exhibit a universal chaotic system. 1.
ContinuousTime Symmetric Hopfield Nets Are Computationally Universal
"... We establish a fundamental result in the theory of computation by continuoustime dynamical systems, by showing that systems corresponding to so called continuoustime symmetric Hopfield nets are capable of general computation. As is well known, such networks have very constrained, Liapunovfunction ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
We establish a fundamental result in the theory of computation by continuoustime dynamical systems, by showing that systems corresponding to so called continuoustime symmetric Hopfield nets are capable of general computation. As is well known, such networks have very constrained, Liapunovfunction controlled dynamics. Nevertheless, we show that they are universal and efficient computational devices, in the sense that any convergent synchronous fully parallel computation by a recurrent network of n discretetime binary neurons, with in general asymmetric coupling weights, can be simulated by a symmetric continuoustime Hopfield net containing only 18n+7 units employing the saturatedlinear activation function. Moreover, if the asymmetric network has maximum integer weight size w_max and converges in discrete time t*, then the corresponding Hopfield net can be designed to operate in continuous time Θ(t*/ε), for any ε > 0...
General Purpose Computation with Neural Networks: A Survey of Complexity Theoretic Results
, 2003
"... We survey and summarize the existing literature on the computational aspects of neural network models, by presenting a detailed taxonomy of the various models according to their complexity theoretic characteristics. The criteria of classi cation include e.g. the architecture of the network (fee ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
We survey and summarize the existing literature on the computational aspects of neural network models, by presenting a detailed taxonomy of the various models according to their complexity theoretic characteristics. The criteria of classi cation include e.g. the architecture of the network (feedforward vs. recurrent), time model (discrete vs. continuous), state type (binary vs. analog), weight constraints (symmetric vs. asymmetric), network size ( nite nets vs. in  nite families), computation type (deterministic vs. probabilistic), etc.
What lies beyond the mountains? Computational systems beyond the Turing limit
 BULLETIN OF THE EUROPEAN ASSOCIATION FOR THEORETICAL COMPUTER SCIENCE
, 2005
"... Up to Turing power, all computations are describable by suitable programs, which correspond to the prescription by finite means of some rational parameters of the system or some computable reals. ¿From Turing power up we have computations that are not describable by finite means: computation without ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Up to Turing power, all computations are describable by suitable programs, which correspond to the prescription by finite means of some rational parameters of the system or some computable reals. ¿From Turing power up we have computations that are not describable by finite means: computation without a program. When we observe natural phenomena and endow them with computational significance, it is not the algorithm we are observing but the process. Some objects near us may be performing hypercomputation: we observe them, but we will never be able to simulate their behaviour on a computer. What is then the profit of such a theory of computation to Science?
Exponential Transients in ContinuousTime Liapunov Systems
 In Proceedings of the 11th ICANN'2001 Conference on Arti Neural Networks, LNCS 2130
, 2003
"... We consider the convergence behavior of a class of continuoustime dynamical systems corresponding to so called symmetric Hopfield nets studied in neural networks theory. We prove that such systems may have transient times that are exponential in the system dimension (i.e. number of "neurons"), desp ..."
Abstract
 Add to MetaCart
We consider the convergence behavior of a class of continuoustime dynamical systems corresponding to so called symmetric Hopfield nets studied in neural networks theory. We prove that such systems may have transient times that are exponential in the system dimension (i.e. number of "neurons"), despite the fact that their dynamics are controlled by Liapunov functions. This result stands in contrast to many proposed uses of such systems in e.g. combinatorial optimization applications, in which it is often implicitly assumed that their convergence is rapid. An additional interesting observation is that our example of an exponentialtransient continuoustime system (a simulated binary counter) in fact converges more slowly than any discretetime Hopfield system of the same representation size. This suggests that continuoustime systems may be worth investigating for gains in descriptional efficiency as compared to their discretetime counterparts.