# A melhor ferramenta para a sua pesquisa, trabalho e TCC!

Página 1 dos resultados de 3079 itens digitais encontrados em 0.028 segundos

## ‣ Corpus-based unit selection for natural-sounding speech synthesis

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 214 p.; 3462128 bytes; 3461885 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
67.397354%

Speech synthesis is an automatic encoding process carried out by machine through which symbols conveying linguistic information are converted into an acoustic waveform. In the past decade or so, a recent trend toward a non-parametric, corpus-based approach has focused on using real human speech as source material for producing novel natural-sounding speech. This work proposes a communication-theoretic formulation in which unit selection is a noisy channel through which an input sequence of symbols passes and an output sequence, possibly corrupted due to the coverage limits of the corpus, emerges. The penalty of approximation is quantified by substitution and concatenation costs which grade what unit contexts are interchangeable and where concatenations are not perceivable. These costs are semi-automatically derived from data and are found to agree with acoustic-phonetic knowledge. The implementation is based on a finite-state transducer (FST) representation that has been successfully used in speech and language processing applications including speech recognition. A proposed constraint kernel topology connects all units in the corpus with associated substitution and concatenation costs and enables an efficient Viterbi search that operates with low latency and scales to large corpora. An A* search can be applied in a second...

Link permanente para citações:

## ‣ Analysis and specification of office procedures

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 232 leaves; 18798434 bytes; 18798193 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
67.464585%

#Electrical Engineering and Computer Science.#Office practice Automation#System analysis#System design#Formal languages#OSL (Computer program language)

by Jay Stuart Kunin.; Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1982.; MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING.; Bibliography: leaves 228-232.

Link permanente para citações:

## ‣ Efficient implementation of applicative languages

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 199 leaves; 10046966 bytes; 10046724 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
77.397354%

by William Beekley Ackerman.; Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1984.; MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING.; Bibliography: leaves 196-199.

Link permanente para citações:

## ‣ Optimal interpreters for lambda-calculus based functional languages

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 197 leaves; 14714672 bytes; 14714431 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
77.397354%

by Vinod Kumar Kathail.; Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1990.; Includes bibliographical references (leaves 195-197).

Link permanente para citações:

## ‣ On the power of nondeterminism in small two-way finite automata; On the power of nondeterminism in small 2DFA

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 26 p.; 1416389 bytes; 1416574 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
67.397354%

We examine the conjecture that one-way nondeterministic finite automata (NFAS) can be exponentially more succinct than two-way deterministic ones (2DFAS); equivalently, that no polynomial-size sequence of 2DFAs can recognize B, for B a particular sequence of regular languages that is among the hardest of those recognizable by polynomial-size sequences of 1NFAs. We prove that the most natural single-pass 2DFA algorithm for deciding B fails, "single-pass" meaning that the automaton is bound to terminate as soon as it reaches an endmarker for the first time. On the way, we introduce the notion of dilemmas as an interesting general tool for constructing hard inputs for 2DFAS.; by Christos Kapoutsis.; Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004.; Includes bibliographical references (p. 25-26).

Link permanente para citações:

## ‣ Event structure and the encoding of arguments : the syntax of the Mandarin and English verb phrase

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 194 p.; 9880232 bytes; 9905049 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
67.397354%

(cont.) to variations in the way functional elements interact with verbal roots. Overall, my work not only contributes to our understanding of how events are syntactically represented, but also explicates interactions at the syntax-semantics interface, clarifying the relationship between surface form, syntactic structure, and logical form. A theory of argument structure grounded in independently-motivated syntactic constraints, on the one hand, and the semantic structure of events, on the other hand, is able to account for a wide range of empirical facts with few stipulations.; This work presents a theory of linguistic representation that attempts to capture the syntactic structure of verbs and their arguments. My framework is based on the assumption that the proper representation of argument structure is event structure. Furthermore, I develop the hypothesis that event structure is syntactic structure, and argue that verb meanings are compositionally derived in the syntax from verbalizing heads, functional elements that license eventive interpretations, and verbal roots, abstract concepts drawn from encyclopedic knowledge. The overall goal of the enterprise is to develop a theory that is able to transparently relate the structure and meaning of verbal arguments. By hypothesis...

Link permanente para citações:

## ‣ A scalable mixed-level approach to dynamic analysis of C and C++ programs

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 112 p.

Português

Relevância na Pesquisa
67.397354%

This thesis addresses the difficult task of constructing robust and scalable dynamic program analysis tools for programs written in memory-unsafe languages such as C and C++, especially those that are interested in observing the contents of data structures at run time. In this thesis, I first introduce my novel mixed-level approach to dynamic analysis, which combines the advantages of both source- and binary-based approaches. Second, I present a tool framework that embodies the mixed-level approach. This framework provides memory safety guarantees, allows tools built upon it to access rich source- and binary-level information simultaneously at run time, and enables tools to scale to large, real-world C and C++ programs on the order of millions of lines of code. Third, I present two dynamic analysis tools built upon my framework - one for performing value profiling and the other for performing dynamic inference of abstract types - and describe how they far surpass previous analyses in terms of scalability, robustness, and applicability. Lastly, I present several case studies demonstrating how these tools aid both humans and automated tools in several program analysis tasks: improving human understanding of unfamiliar code, invariant detection...

Link permanente para citações:

## ‣ Silicon Compilation of Very High Level Languages

Fonte: University of Rochester. Computer Science Department.
Publicador: University of Rochester. Computer Science Department.

Tipo: Thesis; Technical Report

Português

Relevância na Pesquisa
67.719805%

Thesis (Ph. D.)--University of Rochester. Dept. of Computer Science, 1984. Simultaneously published in the Technical Report series.; The report concerns the design and implementation of a compiler for two Very High Level Languages. The first language is a set language similar to VERS or SETL. The second language is a novel signal processing language. The compiler uses data flow and type information to constrain possible choices before choosing a possible implementation. Heuristic search is then used to choose from competing implementations of abstract data types. Constraint propagation is used at every selection step to remove incompatible configurations from the search. Finally, the use of specialized procedures called "design critics" is proposed to resolve global constraint conflicts. The output of the compiler is a parts list, a net list of module interconnections and the fields of the control store.

Link permanente para citações:

## ‣ SUDS : automatic parallelization for raw processors

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 181 p.; 11002412 bytes; 11025390 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
67.616357%

A computer can never be too fast or too cheap. Computer systems pervade nearly every aspect of science, engineering, communications and commerce because they perform certain tasks at rates unachievable by any other kind of system built by humans. A computer system's throughput, however, is constrained by that system's ability to find concurrency. Given a particular target work load the computer architect's role is to design mechanisms to find and exploit the available concurrency in that work load. This thesis describes SUDS (Software Un-Do System), a compiler and runtime system that can automatically find and exploit the available concurrency of scalar operations in imperative programs with arbitrary unstructured and unpredictable control flow. The core compiler transformation that enables this is scalar queue conversion. Scalar queue conversion makes scalar renaming an explicit operation through a process similar to closure conversion, a technique traditionally used to compile functional languages. The scalar queue conversion compiler transformation is speculative, in the sense that it may introduce dynamic memory allocation operations into code that would not otherwise dynamically allocate memory. Thus, SUDS also includes a transactional runtime system that periodically checkpoints machine state...

Link permanente para citações:

## ‣ Starkiller : a static type inferencer and compiler for Python; Static type inferencer for dynamic languages

Fonte: Massachusetts Institute of Technology
Publicador: Massachusetts Institute of Technology

Tipo: Tese de Doutorado
Formato: 96 leaves; 436696 bytes; 435331 bytes; application/pdf; application/pdf

Português

Relevância na Pesquisa
77.397354%

Starkiller is a type inferencer and compiler for the dynamic language Python designed to generate fast native code. It analyzes Python source programs and converts them into equivalent C++ programs. Starkiller's type inference algorithm is based on the Cartesian Product Algorithm but has been significantly modified to support a radically different language. It includes an External Type Description Language that enables extension authors to document how their foreign code extensions interact with Python. This enables Starkiller to analyze Python code that interacts with foreign code written in C, C++, or Fortran. The type inference algorithm also handles data polymorphism in addition to parametric polymorphism, thus improving precision. Starkiller supports the entire Python language except for dynamic code insertion features such as eval and dynamic module loading. While the system is not yet complete, early numeric benchmarks show that Starkiller compiled code performs almost as well as hand made C code and substantially better than alternative Python compilers.; by Michael Salib.; Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004.; Includes bibliographical references (leaves 93-96).; This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.

Link permanente para citações:

## ‣ Log-space counter is useful for unary languages by help of a constant-size quantum register

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
68.01579%

#Computer Science - Computational Complexity#Computer Science - Formal Languages and Automata Theory#Quantum Physics

The minimum amount of resources to recognize a nonregular language is a
fundamental research topic in theoretical computer science which has been
examined for different kinds of resources and many different models. In this
note, we focus on unary languages and space complexity on counters. Our model
is two-way one-counter automaton with quantum and classical states (2QCCA),
which is a two-way finite automaton with one-counter (2DCA) augmented with a
fixed size quantum register or a two-way finite automaton with quantum and
classical states (2QCFA) augmented with a classical counter. It is known that
any 2DCA using a sublinear space on its counter can recognize only regular
languages \cite{DG82B}. In this note, we show that bounded-error 2QCCAs can
recognize a non-regular unary language by using logarithmic space on its
counters for the members. Note that it is still an open problem whether
bounded-error 2QCFA can recognize a non-regular unary language.; Comment: The text is updated by adding a new reference. Technical report. 10
pages. arXiv admin note: text overlap with arXiv:1207.3880

Link permanente para citações:

## ‣ Deterministic pushdown automata and unary languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Publicado em 08/05/2009
Português

Relevância na Pesquisa
67.719805%

The simulation of deterministic pushdown automata defined over a one-letter
alphabet by finite state automata is investigated from a descriptional
complexity point of view. We show that each unary deterministic pushdown
automaton of size s can be simulated by a deterministic finite automaton with a
number of states that is exponential in s. We prove that this simulation is
tight. Furthermore, its cost cannot be reduced even if it is performed by a
two-way nondeterministic automaton. We also prove that there are unary
languages for which deterministic pushdown automata cannot be exponentially
more succinct than finite automata. In order to state this result, we
investigate the conversion of deterministic pushdown automata into context-free
grammars. We prove that in the unary case the number of variables in the
resulting grammar is strictly smaller than the number of variables needed in
the case of nonunary alphabets.; Comment: 17 pages. Preprint of an article submitted for consideration in the
International Journal of Foundations of Computer Science (World Scientific
Publishing Company). A preliminary version was presented at the conference
CIAA 2008

Link permanente para citações:

## ‣ Precedence Automata and Languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
67.82779%

Operator precedence grammars define a classical Boolean and deterministic
context-free family (called Floyd languages or FLs). FLs have been shown to
strictly include the well-known visibly pushdown languages, and enjoy the same
nice closure properties. We introduce here Floyd automata, an equivalent
operational formalism for defining FLs. This also permits to extend the class
to deal with infinite strings to perform for instance model checking.; Comment: Extended version of the paper which appeared in Proceedings of CSR
2011, Lecture Notes in Computer Science, vol. 6651, pp. 291-304, 2011.
Theorem 1 has been corrected and a complete proof is given in Appendix

Link permanente para citações:

## ‣ Proceedings 14th International Conference on Automata and Formal Languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Publicado em 20/05/2014
Português

Relevância na Pesquisa
67.82779%

The 14th International Conference Automata and Formal Languages (AFL 2014)
was held in Szeged, Hungary, from the 27th to the 29th of May, 2014. The
conference was organized by the Department of Foundations of Computer Science
of the University of Szeged. Topics of interest covered the theory and
applications of automata and formal languages and related areas.

Link permanente para citações:

## ‣ Pseudorandom Generators Against Advised Context-Free Languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
68.21298%

#Computer Science - Formal Languages and Automata Theory#Computer Science - Computational Complexity#F.4.3#F.1.1#F.1.3

Pseudorandomness has played a central role in modern cryptography, finding
theoretical and practical applications to various fields of computer science. A
function that generates pseudorandom strings from shorter but truly random
seeds is known as a pseudorandom generator. Our generators are designed to fool
languages (or equivalently, Boolean-valued functions). In particular, our
generator fools advised context-free languages, namely, context-free languages
assisted by external information known as advice, and moreover our generator is
made almost one-to-one, stretching $n$-bit seeds to $n+1$ bits. We explicitly
construct such a pseudorandom generator, which is computed by a deterministic
Turing machine using logarithmic space and also belongs to CFLMV(2)/n---a
functional extension of the 2-conjunctive closure of CFL with the help of
appropriate deterministic advice. In contrast, we show that there is no almost
one-to-one pseudorandom generator against context-free languages if we demand
that it should be computed by a nondeterministic pushdown automaton equipped
with a write-only output tape. Our generator naturally extends known
pseudorandom generators against advised regular languages. Our proof of the
CFL/n-pseudorandomness of the generator is quite elementary...

Link permanente para citações:

## ‣ On Pebble Automata for Data Languages with Decidable Emptiness Problem

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Publicado em 30/10/2009
Português

Relevância na Pesquisa
68.01579%

In this paper we study a subclass of pebble automata (PA) for data languages
for which the emptiness problem is decidable. Namely, we introduce the
so-called top view weak PA. Roughly speaking, top view weak PA are weak PA
where the equality test is performed only between the data values seen by the
two most recently placed pebbles. The emptiness problem for this model is
decidable. We also show that it is robust: alternating, nondeterministic and
deterministic top view weak PA have the same recognition power. Moreover, this
model is strong enough to accept all data languages expressible in Linear
Temporal Logic with the future-time operators, augmented with one register
freeze quantifier.; Comment: An extended abstract of this work has been published in the
proceedings of the 34th International Symposium on Mathematical Foundations
of Computer Science (MFCS) 2009}, Springer, Lecture Notes in Computer Science
5734, pages 712-723

Link permanente para citações:

## ‣ Oracle Pushdown Automata, Nondeterministic Reducibilities, and the CFL Hierarchy over the Family of Context-Free Languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
68.20464%

#Computer Science - Formal Languages and Automata Theory#Computer Science - Computational Complexity

To expand a fundamental theory of context-free languages, we equip
nondeterministic one-way pushdown automata with additional oracle mechanisms,
which naturally induce various nondeterministic reducibilities among formal
languages. As a natural restriction of NP-reducibility, we introduce a notion
of many-one CFL reducibility and conduct a ground work to formulate a coherent
framework for further expositions. Two more powerful reducibilities--bounded
truth-table and Turing CFL-reducibilities--are also discussed in comparison.
The Turing CFL-reducibility, in particular, helps us introduce an exquisite
hierarchy, called the CFL hierarchy, built over the family CFL of context-free
languages. For each level of this hierarchy, its basic structural properties
are proven and three alternative characterizations are presented. The second
level is not included in NC(2) unless NP= NC(2). The first and second levels of
the hierarchy are different. The rest of the hierarchy (more strongly, the
Boolean hierarchy built over each level of the hierarchy) is also infinite
unless the polynomial hierarchy over NP collapses. This follows from a
characterization of the Boolean hierarchy over the k-th level of the polynomial
hierarchy in terms of the Boolean hierarchy over the k+1st level of the CFL
hierarchy using log-space many-one reductions. Similarly...

Link permanente para citações:

## ‣ The separation problem for regular languages by piecewise testable languages

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Publicado em 08/03/2013
Português

Relevância na Pesquisa
67.953765%

Separation is a classical problem in mathematics and computer science. It
asks whether, given two sets belonging to some class, it is possible to
separate them by another set of a smaller class. We present and discuss the
separation problem for regular languages. We then give a direct polynomial time
algorithm to check whether two given regular languages are separable by a
piecewise testable language, that is, whether a $B{\Sigma}1(<)$ sentence can
witness that the languages are indeed disjoint. The proof is a reformulation
and a refinement of an algebraic argument already given by Almeida and the
second author.

Link permanente para citações:

## ‣ Tarski's influence on computer science

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
67.648027%

The influence of Alfred Tarski on computer science was indirect but
significant in a number of directions and was in certain respects fundamental.
Here surveyed is the work of Tarski on the decision procedure for algebra and
geometry, the method of elimination of quantifiers, the semantics of formal
languages, modeltheoretic preservation theorems, and algebraic logic; various
connections of each with computer science are taken up.

Link permanente para citações:

## ‣ Entropy sensitivity of languages defined by infinite automata, via Markov chains with forbidden transitions

Fonte: Universidade Cornell
Publicador: Universidade Cornell

Tipo: Artigo de Revista Científica

Português

Relevância na Pesquisa
67.719805%

#Computer Science - Formal Languages and Automata Theory#Mathematics - Probability#05C63, 37A35, 60J10, 68Q45

A language L over a finite alphabet is growth-sensitive (or entropy
sensitive) if forbidding any set of subwords F yields a sub-language L^F whose
exponential growth rate (entropy) is smaller than that of L. Let (X, E, l) be
an infinite, oriented, labelled graph. Considering the graph as an (infinite)
automaton, we associate with any pair of vertices x,y in X the language
consisting of all words that can be read as the labels along some path from x
to y. Under suitable, general assumptions we prove that these languages are
growth-sensitive. This is based on using Markov chains with forbidden
transitions.; Comment: to appear in Theoretical Computer Science, 2010

Link permanente para citações: