We consider stochastic bandit problems with a continuum set of arms and where the expected re-ward is a continuous and unimodal function of the arm. No further assumption is made regarding thesmoothness and the structure of the expected reward function. We propose Stochastic Pentachotomy(SP), an algorithm for which we derive finite-time regret upper bounds. In particular, we sho...

In previous work we developed a method of learning Bayesian Network models from raw data This method relies on the well known minimal description length MDL principle The MDL principle is particularly well suited to this task as it allows us to tradeo in a principled way the accuracy of the learned network against its practical usefulness In this paper we present some new results that have aris...

In the unidirectional conversion problem an on line player is given the task of converting dollars to yen over some period of time Each day a new exchange rate is announced and the player must decide how many dollars to convert His goal is to minimize the competitive ratio de ned as supE POPT E PX E where E ranges over exchange rate sequences POPT E is the number of yen obtained by an optimal o...

Four out of the 14 second round candidates of the NIST SHA-3 cryptographic hash algo­ rithm competition are so-called AES-inspired algorithms which share common structure and features with AES or even use it as a subroutine. This paper focuses on two of them, Fugue and Grøstl, and studies how efficiently logic can be shared in implementations combining them with AES. It will be shown that addin...

This position paper argues that the Baldwin effectis widely misunderstood by the evolutionarycomputation community. The misunderstandingsappear to fall into two general categories. Firstly,it is commonly believed that the Baldwin effect isconcerned with the synergy that results whenthere is an evolving population of learning indi-viduals. This is only half of the sto...

continuity constraint of motion vectors. Also, many reThis paper proposes a two-stage optical flow estimation algosearchers (e.g., Horn and Schunck [10], Yachida [11], rithm for video sequences based on feature-based matching and Glazer [12], Nagel [13], Nagel and Enkelmann [14], Paquin module-fusion. In the first stage, the algorithm computes initial and Dubois [15], Hildreth [16], Mitiche et ...

We introduce an algorithm that solves the maximum ow problem without generating ows explicitly The algorithm solves di rectly a problem we call the maximum s excess problem That problem is equivalent to the minimum cut problem and is a direct extension of the maximum closure problem The concepts used also lead to a new parametric analysis algorithm generating all breakpoints in the amount of ti...

This paper cons ide rs the following generalization of the We be r plant location probl em: the plant 's output leve l is fix ed , and its levels of input from it s supply point s, as we ll as its location , are among th e decis io n variabl es_ Hurter and W ende ll (J. Reg. Sci. , 1972) s ho wed th at thi s problem ad mits a kind of separabilit y when the plant' s production fun ction li es in...

In k routing each of the n processing units of an n n mesh connected computer initially holds packet which must be routed such that any processor is the destination of at most k packets This problem re ects practical desire for routing better than the popular routing of permutations k routing also has implications for hot potato worm hole routing which is of great importance for real world syst...

This paper presents a new distributed processing ap proach to direct blind equalization of Single Input Multiple Out put SIMO channels Under mild conditions it is shown here that we can recover the original source signal up to its scaled and delayed version by decorrelating the equalizer neural network outputs in spatio temporal domain Spatio temporal anti Hebbian learn ing rule simple local bi...