A simple and fast scheme for code compression for VLIW processors

Summary form only given. A scheme for code compression that has a fast decompression algorithm, which can be implemented using simple hardware, is proposed. The effectiveness of the scheme on the TMS320C62x architecture that includes the overheads of a line address table (LAT) is evaluated and obtained compression rates ranging from 70% to 80%. Two schemes for decompression are proposed. The basic idea underlying the scheme is a simple clustering algorithm that partially maps a block of instructions into a set of clusters. The clustering algorithm is a greedy algorithm based on the frequency of occurrence of various instructions.

Fast algorithms for frequent episode discovery in event sequences

In this paper we consider the process of discovering frequent episodes in event sequences. The most computationally intensive part of this process is that of counting the frequencies of a set of candidate episodes. We present two new frequency counting algorithms for speeding up this part. These, referred to as non-overlapping and non-inteleaved frequency counts, are based on directly counting suitable subsets of the occurrences of an episode. Hence they are different from the frequency counts of Mannila et al [1], where they count the number of windows in which the episode occurs. Our new frequency counts offer a speed-up factor of 7 or more on real and synthetic datasets. We also show how the new frequency counts can be used when the events in episodes have time-durations as well.

An Opportunistic, Fast, and Distributed Subchannel and User-Pairing Algorithm for OFDMA

Channel-aware assignment of subchannels to users in the downlink of an OFDMA system requires extensive feedback of channel state information (CSI) to the base station. Since bandwidth is scarce, schemes that limit feedback are necessary. We develop a novel, low feedback, distributed splitting-based algorithm called SplitSelect to opportunistically assign each subchannel to its most suitable user. SplitSelect explicitly handles multiple access control aspects associated with CSI feedback, and scales well with the number of users. In it, according to a scheduling criterion, each user locally maintains a scheduling metric for each subchannel. The goal is to select, for each subchannel, the user with the highest scheduling metric. At any time, each user contends for the subchannel for which it has the largest scheduling metric among the unallocated subchannels. A tractable asymptotic analysis of a system with many users is central to SplitSelect's simple design. Extensive simulation results demonstrate the speed with which subchannels and users are paired. The net data throughput, when the time overhead of selection is accounted for, is shown to be substantially better than several schemes proposed in the literature. We also show how fairness and user prioritization can be ensured by suitably defining the scheduling metric.

Human action recognition using a fast learning fully complex-valued classifier

In this paper, we use optical flow based complex-valued features extracted from video sequences to recognize human actions. The optical flow features between two image planes can be appropriately represented in the Complex plane. Therefore, we argue that motion information that is used to model the human actions should be represented as complex-valued features and propose a fast learning fully complex-valued neural classifier to solve the action recognition task. The classifier, termed as, ``fast learning fully complex-valued neural (FLFCN) classifier'' is a single hidden layer fully complex-valued neural network. The neurons in the hidden layer employ the fully complex-valued activation function of the type of a hyperbolic secant function. The parameters of the hidden layer are chosen randomly and the output weights are estimated as the minimum norm least square solution to a set of linear equations. The results indicate the superior performance of FLFCN classifier in recognizing the actions compared to real-valued support vector machines and other existing results in the literature. Complex valued representation of 2D motion and orthogonal decision boundaries boost the classification performance of FLFCN classifier. (c) 2012 Elsevier B.V. All rights reserved.

Coexistence of fast photodarkening and slow photobleaching in Ge19As21Se60 thin films

We experimentally demonstrate the coexistence of two opposite photo-effects, viz. fast photodarkening (PD) and slow photobleaching (PB) in Ge19As21Se60 thin films, when illuminated with a laser of wavelength 671 nm. PD appears to begin instantaneously upon light illumination and saturates in tens of seconds. By comparison, PB is a slower process that starts only after PD has saturated. Both PD and PB follow stretched exponetial dependence on time. Modeling of overall change as a linear sum of two contributions suggests that the changes in As and Ge parts of glass network respond to light effectively indepndent of each other. (C) 2012 Optical Society of America

AIWIN - A fast and jigless inspection technique for machined parts

Precision inspection of manufactured components having multiple complex surfaces and variable tolerance definition is an involved, complex and time-consuming function. In routine practice, a jig is used to present the part in a known reference frame to carry out the inspection process. Jigs involve both time and cost in their development, manufacture and use. This paper describes 'as is where is inspection' (AIWIN), a new automated inspection technique that accelerates the inspection process by carrying out a fast registration procedure and establishing a quick correspondence between the part to inspect and its CAD geometry. The main challenge in doing away with a jig is that the inspection reference frame could be far removed from the CAD frame. Traditional techniques based on iterative closest point (ICP) or Newton methods require either a large number of iterations for convergence or fail in such a situation. A two-step coarse registration process is proposed to provide a good initial guess for a modified ICP algorithm developed earlier (Ravishankar et al., Int J Adv Manuf Technol 46(1-4):227-236, 2010). The first step uses a calibrated sphere for local hard registration and fixing the translation error. This transformation locates the centre for the sphere in the CAD frame. In the second step, the inverse transformation (involving pure rotation about multiple axes) required to align the inspection points measured on the manufactured part with the CAD point dataset of the model is determined and enforced. This completes the coarse registration enabling fast convergence of the modified ICP algorithm. The new technique has been implemented on complex freeform machined components and the inspection results clearly show that the process is precise and reliable with rapid convergence. © 2011 Springer-Verlag London Limited.

High field carrier transport in graphene: Insights from fast current transient

In this work, we observe gate tunable negative differential conductance (NDC) and current saturation in single layer and bilayer graphene transistor at high source-drain field, which arise due to the interplay among (1) self-heating, (2) hot carrier injection, and (3) drain induced minority carrier injection. The magnitude of the NDC is found to be reduced for a bilayer, in agreement with its weaker carrier-optical phonon coupling and less efficient hot carrier injection. The contributions of different mechanisms to the observed results are decoupled through fast transient measurements with nanosecond resolution. The findings provide insights into high field transport in graphene. (C) 2012 American Institute of Physics. http://dx.doi.org/10.1063/1.4754103]

Fast image reconstruction for fluorescence microscopy

Real-time image reconstruction is essential for improving the temporal resolution of fluorescence microscopy. A number of unavoidable processes such as, optical aberration, noise and scattering degrade image quality, thereby making image reconstruction an ill-posed problem. Maximum likelihood is an attractive technique for data reconstruction especially when the problem is ill-posed. Iterative nature of the maximum likelihood technique eludes real-time imaging. Here we propose and demonstrate a compute unified device architecture (CUDA) based fast computing engine for real-time 3D fluorescence imaging. A maximum performance boost of 210x is reported. Easy availability of powerful computing engines is a boon and may accelerate to realize real-time 3D fluorescence imaging. Copyright 2012 Author(s). This article is distributed under a Creative Commons Attribution 3.0 Unported License. http://dx.doi.org/10.1063/1.4754604]

As(3)p: A Fast Algorithm to Search Structurally Similar Protiens

Protein structure comparison is essential for understanding various aspects of protein structure, function and evolution. It can be used to explore the structural diversity and evolutionary patterns of protein families. In view of the above, a new algorithm is proposed which performs faster protein structure comparison using the peptide backbone torsional angles. It is fast, robust, computationally less expensive and efficient in finding structural similarities between two different protein structures and is also capable of identifying structural repeats within the same protein molecule.

Feedback Overhead-Aware, Distributed, Fast, and Reliable Selection

In a communication system in which K nodes communicate with a central sink node, the following problem of selection often occurs. Each node maintains a preference number called a metric, which is not known to other nodes. The sink node must find the `best' node with the largest metric. The local nature of the metrics requires the selection process to be distributed. Further, the selection needs to be fast in order to increase the fraction of time available for data transmission using the selected node and to handle time-varying environments. While several selection schemes have been proposed in the literature, each has its own shortcomings. We propose a novel, distributed selection scheme that generalizes the best features of the timer scheme, which requires minimal feedback but does not guarantee successful selection, and the splitting scheme, which requires more feedback but guarantees successful selection. The proposed scheme introduces several new ideas into the design of the timer and splitting schemes. It explicitly accounts for feedback overheads and guarantees selection of the best node. We analyze and optimize the performance of the scheme and show that it is scalable, reliable, and fast. We also present new insights about the optimal timer scheme.

SURE-FAST BILATERAL FILTERS

Edge-preserving smoothing is widely used in image processing and bilateral filtering is one way to achieve it. Bilateral filter is a nonlinear combination of domain and range filters. Implementing the classical bilateral filter is computationally intensive, owing to the nonlinearity of the range filter. In the standard form, the domain and range filters are Gaussian functions and the performance depends on the choice of the filter parameters. Recently, a constant time implementation of the bilateral filter has been proposed based on raisedcosine approximation to the Gaussian to facilitate fast implementation of the bilateral filter. We address the problem of determining the optimal parameters for raised-cosine-based constant time implementation of the bilateral filter. To determine the optimal parameters, we propose the use of Stein's unbiased risk estimator (SURE). The fast bilateral filter accelerates the search for optimal parameters by faster optimization of the SURE cost. Experimental results show that the SURE-optimal raised-cosine-based bilateral filter has nearly the same performance as the SURE-optimal standard Gaussian bilateral filter and the Oracle mean squared error (MSE)-based optimal bilateral filter.

Fast Likelihood Computation in Speech Recognition using Matrices

Acoustic modeling using mixtures of multivariate Gaussians is the prevalent approach for many speech processing problems. Computing likelihoods against a large set of Gaussians is required as a part of many speech processing systems and it is the computationally dominant phase for Large Vocabulary Continuous Speech Recognition (LVCSR) systems. We express the likelihood computation as a multiplication of matrices representing augmented feature vectors and Gaussian parameters. The computational gain of this approach over traditional methods is by exploiting the structure of these matrices and efficient implementation of their multiplication. In particular, we explore direct low-rank approximation of the Gaussian parameter matrix and indirect derivation of low-rank factors of the Gaussian parameter matrix by optimum approximation of the likelihood matrix. We show that both the methods lead to similar speedups but the latter leads to far lesser impact on the recognition accuracy. Experiments on 1,138 work vocabulary RM1 task and 6,224 word vocabulary TIMIT task using Sphinx 3.7 system show that, for a typical case the matrix multiplication based approach leads to overall speedup of 46 % on RM1 task and 115 % for TIMIT task. Our low-rank approximation methods provide a way for trading off recognition accuracy for a further increase in computational performance extending overall speedups up to 61 % for RM1 and 119 % for TIMIT for an increase of word error rate (WER) from 3.2 to 3.5 % for RM1 and for no increase in WER for TIMIT. We also express pairwise Euclidean distance computation phase in Dynamic Time Warping (DTW) in terms of matrix multiplication leading to saving of approximately of computational operations. In our experiments using efficient implementation of matrix multiplication, this leads to a speedup of 5.6 in computing the pairwise Euclidean distances and overall speedup up to 3.25 for DTW.

On the Compatibility of a Flux Transport Dynamo with a Fast Tachocline Scenario

The compatibility of the fast-tachocline scenario with a flux-transport dynamo model is explored. We employ a flux-transport dynamo model coupled with simple feedback formulae relating the thickness of the tachocline to the amplitude of the magnetic field or to the Maxwell stress. The dynamo model is found to be robust against the nonlinearity introduced by this simplified fast-tachocline mechanism. Solar-like butterfly diagrams are found to persist and, even without any parameter fitting, the overall thickness of the tachocline is well within the range admitted by helioseismic constraints. In the most realistic case of a time-and latitude-dependent tachocline thickness linked to the value of the Maxwell stress, both the thickness and its latitudinal dependence are in excellent agreement with seismic results. In nonparametric models, cycle-related temporal variations in tachocline thickness are somewhat larger than admitted by helioseismic constraints; we find, however, that introducing a further parameter into our feedback formula readily allows further fine tuning of the thickness variations.