At+t Corp.

At&t Intellectual Property I, L.p.

Lucent Technologies Inc.

Adam Louis Buchsbaum

Raffaele Giancarlo

Jeffery Rex Westbrook

Balachander Krishnamurthy

Jia Wang

Kiem-Phong Vo

Glenn Stephen Fowler

Jan Pieter VanSanten

A preload library partitions certain files into segments and transforms the data of the files on a segment by segment basis, each independently of the other. The transformed segments are then stored as part of a transformed file, with each segment corresponding to a bag containing a chunk (of transformed data) and a gap which permits the chunks to grow when data of the segment is modified during a write operation. The bags are followed by a chunk map which contains metadata defining the compressed version of each segment. The preload library intercepts system calls from applications to the OC specifying access to specific segments of a transformed file. The preload library identifies the chunks corresponding to the specified segments, retrieves and reverses the transform on those chunks and passes the data to the application requesting it. For a system call to write data to certain segments, the data is modified and re-transformed by the library before being written back to disk. A preload library provides improved random access into transformed (e.g. compressed, encrypted, etc.) files.

Data storage technique

A method for clustering together network IP addresses is disclosed. A number of IP addresses are received and processed to determine which IP addresses share a longest prefix matching. The longest prefix matching process is performed according to radix encoded trie which facilitates on-line clustering of the IP addresses. Client and/or server IP addresses may be clustered in accordance with the teachings herein.

Method for fast network-aware clustering

Preload library for transparent file transformation

A system and method of determining a common neighborhood of users sharing a common activity from a plurality of users is provided. The system and method may be used to predict, for a user in the common neighborhood of users, a potential activity from the activities of at least one other user in the common neighborhood of users.

Graph algorithm for common neighborhood analysis

The present invention increases the efficiency of performing longest prefix matching operations by selecting a radix-encoded trie structure optimized with respect to memory cost. The structure is optimized by determining memory costs for retrie structures indexed on different numbers of high-order characters, and then selecting the structure corresponding to the lowest memory cost. The optimization improves performance in IP look-up operations as well as longest-prefix matching operations performed on general alphabets.

Fast prefix matching of bounded strings

An approximate weighted finite-state automaton can be constructed in place of a weighted finite-state automaton so long as the approximate weighted finite-state automaton maintains a sufficient portion of the original best strings in the weighted finite-state automaton and sufficiently few spurious strings are introduced into the approximate weighted finite-state automaton compared to the weighted finite-state automaton. An approximate weighted finite-state automaton can be created from a non-deterministic weighted finite-state automaton during determinization by discarding the requirement that old states be used in place of new states only when an old state is identical to a new state. Instead, in an approximate weighted finite-state automaton, old states will be used in place of new states when each of the remainders of the new state is sufficiently close to the corresponding remainder of the old state. An error tolerance parameter &tgr; defines whether the remainders of the new state are sufficiently close to the corresponding remainders of the old state. If the remainders of the new state are sufficiently close to the remainders of the old state, a new transition is created from a current state to the old state rather than the new state. Such approximate weighted finite-state automata allow the size of the resulting deterministic finite-state automata to be reduced and can prevent the expansion that would otherwise occur in some deterministic finite-state automata.

Method and apparatus for generating deterministic approximate weighted finite-state automata

An approximate weighted finite-state automaton can be constructed in place of a weighted finite-state automaton so long as the approximate weighted finite-state automaton maintains a sufficient portion of the original best strings in the weighted finite-state automaton and sufficiently few spurious strings are introduced into the approximate weighted finite-state automaton compared to the weighted finite-state automaton. An approximate weighted finite-state automaton can be created from a non-deterministic weighted finite-state automaton during determinization by discarding the requirement that old states be used in place of new states only when an old state is identical to a new state. Instead, in an approximate weighted finite-state automaton, old states will be used in place of new states when each of the remainders of the new state is sufficiently close to the corresponding remainder of the old state. An error tolerance parameter .tau. defines whether the remainders of the new state are sufficiently close to the corresponding remainders of the old state. If the remainders of the new state are sufficiently close to the remainders of the old state, a new transition is created from a current state to the old state rather than the new state. Such approximate weighted finite-state automata allow the size of the resulting deterministic finite-state automata to be reduced and can prevent the expansion that would otherwise occur in some deterministic finite-state automata.

Method and apparatus for generating deterministic approximate weighted

A system and method are described for determining a near-optimum subset of data, based on a selected model, from a large corpus of data. Sets of feature vectors corresponding to natural or other preselected divisions of the data corpus are mapped into matrices representative of such divisions. The invention operates to find a submatrix of full rank formed as a union of one or more of those division-based matrices. A greedy algorithm utilizing Gram-Schmidt orthonormalization operates on the division matrices to find a near optimum submatrix and in a time bound representing a substantial improvement over prior-art methods. An important application of the invention is the selection of a small number of sentences from a corpus of a very large number of such sentences from which the parameters of a duration model for speech synthesis can be estimated.

Growing community of inventors

Madison, NJ, United States of America

Adam Louis Buchsbaum