Filters








57,160 Hits in 3.3 sec

Capacity and capacity-achieving input distribution of the energy detector

Erik Leitinger, Bernhard C. Geiger, Klaus Witrisal
2012 2012 IEEE International Conference on Ultra-Wideband  
A proper statistical model is introduced which makes it possible to treat the energy detector as a constrained continuous communication channel.  ...  This paper presents the capacity-achieving input distribution of an energy detection receiver structure.  ...  The algorithm starts with an initial list of mass points P Figure 3 shows the capacity for different M values over a wide range of E s /N 0 .  ... 
doi:10.1109/icuwb.2012.6340409 dblp:conf/icuwb/LeitingerGW12 fatcat:x37ukotvyvg4pk5sm5bl3qh23a

Guest Editors' Introduction: Special issue on Learning Theory (COLT-2007)

Nader H. Bshouty, Claudio Gentile
2008 Machine Learning  
In this sequential setting, the learning algorithm is asked to compute a distribution over a set of experts, and receives a gain equal to the average instantaneous gain over the experts.  ...  (iii) A hardness result for learning circuits with large alphabets when there are no restrictions on their topologies.  ... 
doi:10.1007/s10994-008-5062-x fatcat:zvcvezl4tfdlheiudd5gxrqbea

Composition of weighted finite transducers in MapReduce

Bilal Elghadyry, Faissal Ouardi, Sébastien Verel
2021 Journal of Big Data  
The NP-hardness of the composition computation problem presents a challenge that leads us to devise efficient algorithms on a large scale when considering more than two transducers.  ...  Finally, intensive experiments on a wide range of weighted finite-state transducers are conducted to compare the proposed methods and show their efficiency for large-scale data.  ...  In this section, we will focus on how distributed computing program works over the Hadoop MapReduce Model. First, the Hadoop framework and the MapReduce programming model will be briefly presented.  ... 
doi:10.1186/s40537-020-00397-4 fatcat:ltelafcd4zbq5nlaszy46uwkgm

PhD Dissertation: Generalized Independent Components Analysis Over Finite Alphabets [article]

Amichai Painsky
2018 arXiv   pre-print
ICA over finite fields is a special case of ICA in which both the observations and the independent components are over a finite alphabet.  ...  Usually the ICA framework assumes a model according to which the observations are generated (such as a linear transformation with additive noise).  ...  Thank you for taking part in this journey with me.  ... 
arXiv:1809.05043v4 fatcat:qo5cl7ui7zeetcpijrg43skkli

Discovering a domain alphabet

Michael D. Schmidt, Hod Lipson
2009 Proceedings of the 11th Annual conference on Genetic and evolutionary computation - GECCO '09  
An alphabet that is too granular may generate an unnecessarily large search space; an inappropriately coarse grained alphabet may bias or prevent finding optimal solutions.  ...  Here we introduce a method that automatically identifies a small alphabet for a problem domain.  ...  Rather than having to re-derive common terms from scratch, over and over again for each model, the algorithm could benefit from the coarser search of assembling higher order building blocks.  ... 
doi:10.1145/1569901.1570047 dblp:conf/gecco/SchmidtL09 fatcat:l5gnoep4rnh2dp3wmu5g2brkwy

Large alphabets: Finite, infinite, and scaling models

Mesrob I. Ohannessian, Munther A. Dahleh
2012 2012 46th Annual Conference on Information Sciences and Systems (CISS)  
In this paper we present an overview of models for large alphabets, some recent results, and possible directions in this area.  ...  How can we effectively model situations with large alphabets?  ...  SCALING MODELS Finite models for large alphabets, such as the ones we presented, and which are described as uniform consistency over classes of finite sources are adequate for directly capturing alphabet  ... 
doi:10.1109/ciss.2012.6310941 dblp:conf/ciss/OhannessianD12 fatcat:gwhz7ahgrfhxlecdbl5nefoo3u

Learning from Uncertain Data [chapter]

Mehryar Mohri
2003 Lecture Notes in Computer Science  
for computing them efficiently.  ...  But, to deal with recent problems arising in this field, machine learning techniques must be generalized to deal with uncertain data, or datasets whose elements are distributions over sequences, such as  ...  More 1 Many modeling algorithms can be naturally generalized to deal with input distributions by replacing the quantity X originally derived from text by its expectation E[X] based on the probability distribution  ... 
doi:10.1007/978-3-540-45167-9_48 fatcat:q3sqeiremzcgdiozz5sovakdjq

An MCMC Approach to Lossy Compression of Continuous Sources

Dror Baron, Tsachy Weissman
2010 2010 Data Compression Conference  
However, the large alphabet slows down the convergence to the RD function, and is thus an impediment in practice.  ...  We thus propose an MCMC-based algorithm that uses a (smaller) adaptive reproduction alphabet.  ...  In order to approach the RD function closely, Y may need to be large. Not only does the reproduction alphabet become large only for large n, but the large Y slows down the algorithm.  ... 
doi:10.1109/dcc.2010.11 dblp:conf/dcc/BaronW10 fatcat:cbg2gqghs5cspkxq2dmm74lwoa

On the Expected Sublinearity of the Boyer–Moore Algorithm

R. Schaback
1988 SIAM journal on computing (Print)  
with an arbitrary distribution of probabilities.  ...  This paper analyzes the expected performance of a simpli ed version BM 0 of the Boyer{Moore string matching algorithm.  ...  Then the randomized algorithm will model the behavior of the deterministic algorithm quite well even for natural languages, since natural language characters sampled over large intervals can be considered  ... 
doi:10.1137/0217041 fatcat:i6yfd4fuyjhkxl37xxe32ldp3i

Mining Statistically Significant Substrings using the Chi-Square Statistic [article]

Mayank Sachan, Arnab Bhattacharya
2012 arXiv   pre-print
Given a string of characters generated from a memoryless Bernoulli model, the problem is to identify the substring for which the empirical distribution of single letters deviates the most from the distribution  ...  We also describe some applications of our algorithm on cryptology and real world data from finance and sports. Finally, we compare our technique with the existing heuristics for finding the MSS.  ...  Figure 1a depicts the comparison of number of iterations required by our algorithm visa-vis the trivial algorithm for input strings of different lengths (n) generated from the null model for an alphabet  ... 
arXiv:1207.0144v1 fatcat:tqfgmiogy5eajapzdfdqrv2q5m

A Spectral Learning Algorithm for Finite State Transducers [chapter]

Borja Balle, Ariadna Quattoni, Xavier Carreras
2011 Lecture Notes in Computer Science  
At its core, the algorithm is simple, and scalable to large data sets. We present experiments that validate the effectiveness of the algorithm on synthetic and real data.  ...  Recently, Hsu et al. [13] proposed a spectral method for learning Hidden Markov Models (HMMs) which is based on an Observable Operator Model (OOM) view of HMMs.  ...  First, we randomly selected a distribution over input sequences of length three, for input alphabet sizes ranging from 2 to 10, and choosing among uniform, gaussian and power distributions with random  ... 
doi:10.1007/978-3-642-23780-5_20 fatcat:gkwrskycynd6lmjitdg3xa2x7q

Using Substitution Matrices to Estimate Probability Distributions for Biological Sequences

Eleazar Eskin, William Stafford Noble, Yoram Singer
2002 Journal of Computational Biology  
In this paper we present a biologically-motivated method for estimating probability distributions over discrete alphabets from observations using a mixture model of common ancestors.  ...  In contrast to previous such methods, our method has a simple Bayesian interpretation and has the advantage over Dirichlet mixtures that it is both effective and simple to compute for large alphabets.  ...  Leslie for useful comments.  ... 
doi:10.1089/10665270260518263 pmid:12614546 fatcat:27emwo2itzcojf6om6yrqnnlza

EXIT chart approximations using the role model approach

Jossy Sayir
2010 2010 IEEE International Symposium on Information Theory  
We propose an approximation we call mixed information that constitutes a lower bound for the true EXIT function and can be estimated by statistical methods even when the message alphabet is large and histogram-based  ...  Extrinsic Information Transfer (EXIT) functions can be measured by statistical methods if the message alphabet size is moderate or if messages are true a-posteriori distributions.  ...  ACKNOWLEDGMENTS The author is indebted to his office mate Simon Hill for his help in clarifying the relation between the role model approach and Monte Carlo integration, and hopes that this acknowledgment  ... 
doi:10.1109/isit.2010.5513589 dblp:conf/isit/Sayir10 fatcat:jegxaf64hved7ayioaglcyxkfq

A unified framework for the specification and run-time detection of dynamic properties in distributed computations

Özalp Babaoğlu, Eddy Fromentin, Michel Raynal
1996 Journal of Systems and Software  
Considering boolean predicates on states of the computation as an alphabet, dynamic property specification is akin to defining a language over this alphabet.  ...  This formal language-oriented view not only unifies a large body of work on distributed debugging and property detection, it also leads to simple and efficient detection algorithms.  ...  Plouzeau for interesting discussions related to the debugging of distributed programs.  ... 
doi:10.1016/0164-1212(96)00027-1 fatcat:fqlyqqcaezhvzihwyqwik4pxci

Fast dictionary attacks on passwords using time-space tradeoff

Arvind Narayanan, Vitaly Shmatikov
2005 Proceedings of the 12th ACM conference on Computer and communications security - CCS '05  
Our second contribution is an algorithm for efficient enumeration of the remaining password space.  ...  This is a much higher percentage than Oechslin's "rainbow" attack, which is the fastest currently known technique for searching large keyspaces.  ...  We are grateful to Dmitry Sumin of Passware for providing the password material for our experiments.  ... 
doi:10.1145/1102120.1102168 dblp:conf/ccs/NarayananS05a fatcat:phyy3i6gmffq7er5evgwhcvgle
« Previous Showing results 1 — 15 out of 57,160 results