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.

Maintaining information in fully dynamic trees with top trees

Abstract: We design top trees as a new simpler interface for data structures maintaining information in a fully dynamic forest. We demonstrate how easy and versatile they are to use on a host of different applications. For example, we show how to maintain the diameter, center, and median of each tree in the forest. The forest can be updated by insertion and deletion of edges and by changes to vertex and edge weights. Each update is supported in O(log n) time, where n is the size of the tree(s) involved in the update. Also, we show how to support nearest common ancestor queries and level ancestor queries with respect to arbitrary roots in O(log n) time. Finally, with marked and unmarked vertices, we show how to compute distances to a nearest marked vertex. The latter has applications to approximate nearest marked vertex in general graphs, and thereby to static optimization problems over shortest path metrics.Technically speaking, top trees are easily implemented either with Frederickson's [1997a] topology trees or with Sleator and Tarjan's [1983] dynamic trees. However, we claim that the interface is simpler for many applications, and indeed our new bounds are quadratic improvements over previous bounds where they exist.

Multi-band speech recognition in noisy environments

Abstract: This paper presents a new approach for multi-band based automatic speech recognition (ASR). Previous work by Bourlard et al. (see Proc. Int. Conf. on Spoken Language Processing, Philadelphia, p.426-9, 1996) and Hermansky et al. (see Proc. Int. Conf. on Spoken Language Processing, Philadelphia, p.1579-82, 1996) suggests that multi-band ASR gives a more accurate recognition, especially in noisy acoustic environments, by combining the likelihoods of different frequency bands. Here we evaluate this likelihood recombination (LC) approach to multi-band ASR, and propose an alternative method, namely feature recombination (FC). In the FC system, after different acoustic analyzers are applied to each sub-band individually, a vector is composed by combining the sub-band features. The speech classifier then calculates the likelihood from the single vector. Thus, band-limited noise affects only a few of the feature components, as in the multi-band LC system, but, at the same time, all feature components are jointly modeled, as in conventional ASR. The experimental results show that the FC system can yield better performance than both the conventional ASR and the LC strategy for noisy speech.

Random evolution in massive graphs

Abstract: Many massive graphs (such as the WWW graph and Call graphs) share certain universal characteristics which can be described by the so-called "power law." In this paper, we, examine three important aspects of power law graphs, (1) the evolution of power law graphs, (2) the asymmetry of in-degrees and out-degrees, (3) the "scale invariance" of power law graphs. In particular, we give three increasingly general directed graph models and one general undirected graph model for generating power law graphs by adding at most one node and possibly one or more edges at a time. We show that for any given edge density and desired power laws for in-degrees and out-degrees, not necessarily the same, the resulting graph will almost surely have the desired edge density and the power laws for the in-degrees and out-degrees. Our most general directed and undirected models include nearly all known power law evolution models as special cases. Finally, we show that our evolution models generate "scale invariant" graphs. We describe a method for scaling the time in our evolution model such that the power law of the degree sequences remains invariant.

Adaptive frequency-domain equalization and diversity combining for broadband wireless communications

Abstract: We introduce a new kind of adaptive equalizer that operates in the spatial-frequency domain, and uses either least mean square (LMS) or recursive least squares (RLS) adaptive processing. We simulate the equalizer's performance in an 8 Mb/s QPSK (quaternary phase shift keying) link over a frequency-selective, Rayleigh fading multipath channel with /spl sim/3 /spl mu/s RMS delay spread, corresponding to 60 symbols of dispersion. Our results show rapid convergence and tracking for a range of mobile speeds (up to /spl sim/100 mph). Moreover, a 2-branch RLS equalizer requires only /spl sim/50 complex operations per detected bit, which, at 8 Mb/s, is close to achievable with state-of-the-art digital signal processing technology.

Cancellation of timing and amplitude jitter in symmetric links using highly dispersed pulses

Abstract: We study the properties of symmetric dispersion compensation in optical links using highly dispersed pulse transmission. We show analytically that by splitting the dispersion compensation equally between the input and output of the link, complete cancellation of the timing and amplitude jitter can be obtained in systems where the power profile is symmetric about the center. We explain the dynamics of this cancellation and show, theoretically and experimentaily, that with practical system parameters, symmetric compensation may lead to a considerable improvement in performance.