AT&T Labs

About: AT&T Labs is a based out in . It is known for research contribution in the topics: Network packet & The Internet. The organization has 1879 authors who have published 5595 publications receiving 483151 citations.

A Two-Round Variant of EM for Gaussian Mixtures

Abstract: We show that, given data from a mixture of k well-separated spherical Gaussians in Rn, a simple two-round variant of EM will, with high probability, learn the centers of the Gaussians to near-optimal precision, if the dimension is high (n ≫ log k). We relate this to previous theoretical and empirical work on the EM algorithm.

Local search with perturbations for the prize‐collecting Steiner tree problem in graphs

Abstract: Given an undirected graph with prizes associated with its nodes and weights associated with its edges, the prize-collecting Steiner tree problem consists of finding a subtree of this graph which minimizes the sum of the weights of its edges plus the prizes of the nodes not spanned. In this paper, we describe a multistart local search algorithm for the prize-collecting Steiner tree problem, based on the generation of initial solutions by a primal-dual algorithm using perturbed node prizes. Path-relinking is used to improve the solutions found by local search and variable neighborhood search is used as a post-optimization procedure. Computational experiments involving different algorithm variants are reported. Our results show that the local search with perturbations approach found optimal solutions on nearly all of the instances tested. © 2001 John Wiley & Sons, Inc.

64-Tb/s, 8 b/s/Hz, PDM-36QAM Transmission Over 320 km Using Both Pre- and Post-Transmission Digital Signal Processing

Abstract: We report the successful transmission of 64 Tb/s capacity (640 ×107 Gb/s with 12.5 GHz channel spacing) over 320 km reach utilizing 8-THz of spectrum in the C+L -bands at a net spectral efficiency of 8 bit/s/Hz. Such a result is accomplished by the use of raised-cosine pulse-shaped PDM-36QAM modulation, intradyne detection, both pre- and post-transmission digital equalization, and ultra-large-area fiber. We discuss in detail the digital modulation technology and signal processing algorithms used in the experiment, including a new two-stage, blind frequency-search-based frequency-offset estimation algorithm and a more computationally efficient carrier-phase recovery algorithm.

One-pass wavelet decompositions of data streams

Abstract: We present techniques for computing small space representations of massive data streams. These are inspired by traditional wavelet-based approximations that consist of specific linear projections of the underlying data. We present general "sketch"-based methods for capturing various linear projections and use them to provide pointwise and rangesum estimation of data streams. These methods use small amounts of space and per-item time while streaming through the data and provide accurate representation as our experiments with real data streams show.