Divide and Conquer: Efficient Large-Scale Structure from Motion Using Graph Partitioning

Despite significant advances in recent years, structure-from-motion (SfM) pipelines suffer from two important drawbacks. Apart from requiring significant computational power to solve the large-scale computations involved, such pipelines sometimes fail to correctly reconstruct when the accumulated error in incremental reconstruction is large or when the number of 3D to 2D correspondences are insufficient. In this paper we present a novel approach to mitigate the above-mentioned drawbacks. Using an image match graph based on matching features we partition the image data set into smaller sets or components which are reconstructed independently. Following such reconstructions we utilise the available epipolar relationships that connect images across components to correctly align the individual reconstructions in a global frame of reference. This results in both a significant speed up of at least one order of magnitude and also mitigates the problems of reconstruction failures with a marginal loss in accuracy. The effectiveness of our approach is demonstrated on some large-scale real world data sets.

A Randomized Algorithm for Large Scale Support Vector Learning

We propose a randomized algorithm for large scale SVM learning which solves the problem by iterating over random subsets of the data. Crucial to the algorithm for scalability is the size of the subsets chosen. In the context of text classification we show that, by using ideas from random projections, a sample size of O(log n) can be used to obtain a solution which is close to the optimal with a high probability. Experiments done on synthetic and real life data sets demonstrate that the algorithm scales up SVM learners, without loss in accuracy. 1

Integrated parallelization of computations and visualization for large-scale applications

Critical applications like cyclone tracking and earthquake modeling require simultaneous high-performance simulations and online visualization for timely analysis. Faster simulations and simultaneous visualization enable scientists provide real-time guidance to decision makers. In this work, we have developed an integrated user-driven and automated steering framework that simultaneously performs numerical simulations and efficient online remote visualization of critical weather applications in resource-constrained environments. It considers application dynamics like the criticality of the application and resource dynamics like the storage space, network bandwidth and available number of processors to adapt various application and resource parameters like simulation resolution, simulation rate and the frequency of visualization. We formulate the problem of finding an optimal set of simulation parameters as a linear programming problem. This leads to 30% higher simulation rate and 25-50% lesser storage consumption than a naive greedy approach. The framework also provides the user control over various application parameters like region of interest and simulation resolution. We have also devised an adaptive algorithm to reduce the lag between the simulation and visualization times. Using experiments with different network bandwidths, we find that our adaptive algorithm is able to reduce lag as well as visualize the most representative frames.

Channel Hardening-Exploiting Message Passing (CHEMP) Receiver in Large-Scale MIMO Systems

In this paper, we propose a multiple-input multiple-output (MIMO) receiver algorithm that exploits channel hardening that occurs in large MIMO channels. Channel hardening refers to the phenomenon where the off-diagonal terms of the matrix become increasingly weaker compared to the diagonal terms as the size of the channel gain matrix increases. Specifically, we propose a message passing detection (MPD) algorithm which works with the real-valued matched filtered received vector (whose signal term becomes, where is the transmitted vector), and uses a Gaussian approximation on the off-diagonal terms of the matrix. We also propose a simple estimation scheme which directly obtains an estimate of (instead of an estimate of), which is used as an effective channel estimate in the MPD algorithm. We refer to this receiver as the channel hardening-exploiting message passing (CHEMP) receiver. The proposed CHEMP receiver achieves very good performance in large-scaleMIMO systems (e.g., in systems with 16 to 128 uplink users and 128 base station antennas). For the considered large MIMO settings, the complexity of the proposed MPD algorithm is almost the same as or less than that of the minimum mean square error (MMSE) detection. This is because the MPD algorithm does not need a matrix inversion. It also achieves a significantly better performance compared to MMSE and other message passing detection algorithms using MMSE estimate of. Further, we design optimized irregular low density parity check (LDPC) codes specific to the considered large MIMO channel and the CHEMP receiver through EXIT chart matching. The LDPC codes thus obtained achieve improved coded bit error rate performance compared to off-the-shelf irregular LDPC codes.

Large-Scale Multiuser SM-MIMO Versus Massive MIMO

Spatial modulation (SM) is attractive for multiantenna wireless communications. SM uses multiple transmit antenna elements but only one transmit radio frequency (RF) chain. In SM, in addition to the information bits conveyed through conventional modulation symbols (e.g., QAM), the index of the active transmit antenna also conveys information bits. In this paper, we establish that SM has significant signal-to-noise (SNR) advantage over conventional modulation in large-scale multiuser (multiple-input multiple-output) MIMO systems. Our new contribution in this paper addresses the key issue of large-dimension signal processing at the base station (BS) receiver (e.g., signal detection) in large-scale multiuser SM-MIMO systems, where each user is equipped with multiple transmit antennas (e.g., 2 or 4 antennas) but only one transmit RF chain, and the BS is equipped with tens to hundreds of (e.g., 128) receive antennas. Specifically, we propose two novel algorithms for detection of large-scale SM-MIMO signals at the BS; one is based on message passing and the other is based on local search. The proposed algorithms achieve very good performance and scale well. For the same spectral efficiency, multiuser SM-MIMO outperforms conventional multiuser MIMO (recently being referred to as massive MIMO) by several dBs. The SNR advantage of SM-MIMO over massive MIMO can be attributed to: (i) because of the spatial index bits, SM-MIMO can use a lower-order QAM alphabet compared to that in massive MIMO to achieve the same spectral efficiency, and (ii) for the same spectral efficiency and QAM size, massive MIMO will need more spatial streams per user which leads to increased spatial interference.

Hydroclimatic influence of large-scale circulation on the variability of reservoir inflow

In this study, the nature of basin-scale hydroclimatic association for Indian subcontinent is investigated. It is found that, the large-scale circulation information from Indian Ocean is also equally important in addition to the El Nino-Southern Oscillation (ENSO), owing to the geographical location of Indian subcontinent. The hydroclimatic association of the variation of monsoon inflow into the Hirakud reservoir in India is investigated using ENSO and EQUatorial INdian Ocean Oscillation (EQUINOO, the atmospheric part of Indian Ocean Dipole mode) as the large-scale circulation information from tropical Pacific Ocean and Indian Ocean regions respectively. Individual associations of ENSO & EQUINOO indices with inflow into Hirakud reservoir are also assessed and found to be weak. However, the association of inflows into Hirakud reservoir with the composite index (CI) of ENSO and EQUINOO is quite strong. Thus, the large-scale circulation information from Indian Ocean is also important apart form the ENSO. The potential of the combined information of ENSO and EQUINOO for predicting the inflows during monsoon is also investigated with promising results. The results of this study will be helpful to water resources managers due to fact that the nature of monsoon inflow is becoming available as an early prediction.

The sensitivity of the northern hemisphere summer mean meridional circulation to changes in the large scale eddy forcing

A zonally averaged version of the Goddard Laboratory for Atmospheric Sciences (GLAS) climate model is used to study the sensitivity of the northern hemisphere (NH) summer mean meridional circulation to changes in the large scale eddy forcing. A standard solution is obtained by prescribing the latent heating field and climatological horizontal transports of heat and momentum by the eddies. The radiative heating and surface fluxes are calculated by model parameterizations. This standard solution is compared with the results of several sensitivity studies. When the eddy forcing is reduced to 0.5 times or increased to 1.5 times the climatological values, the strength of the Ferrel cells decrease or increase proportionally. It is also seen that such changes in the eddy forcing can influence the strength of theNH Hadley cell significantly. Possible impact of such changes in the large scale eddy forcing on the monsoon circulation via changes in the Hadley circulation is discussed. Sensitivity experiments including only one component of eddy forcing at a time show that the eddy momentum fluxes seem to be more important in maintaining the Ferrel cells than the eddy heat fluxes. In the absence of the eddy heat fluxes, the observed eddy momentum fluxes alone produce subtropical westerly jets which are weaker than those in the standard solution. On the other hand, the observed eddy heat fluxes alone produce subtropical westerly jets which are stronger than those in the standard solution.

A simple method for large-scale preparation of chick embryo tRNA

A simple method for preparing bulk quantities of tRNA from chick embryo has been developed. In this method chick embryos were homogenized in a buffer of pH 4.5, followed by deproteinization with phenol. The aqueous layer was allowed to separate under gravity. The resulting aqueous layer, after two more phenol treatments, was directly passed through a DEAE-cellulose column and the tRNA eluted therefrom with 1 Image NaCl. The tRNA prepared by this method was as active as the one prepared at neutral pH.

An FETI-preconditioned conjuerate gradient method for large-scale stochastic finite element problems

In the spectral stochastic finite element method for analyzing an uncertain system. the uncertainty is represented by a set of random variables, and a quantity of Interest such as the system response is considered as a function of these random variables Consequently, the underlying Galerkin projection yields a block system of deterministic equations where the blocks are sparse but coupled. The solution of this algebraic system of equations becomes rapidly challenging when the size of the physical system and/or the level of uncertainty is increased This paper addresses this challenge by presenting a preconditioned conjugate gradient method for such block systems where the preconditioning step is based on the dual-primal finite element tearing and interconnecting method equipped with a Krylov subspace reusage technique for accelerating the iterative solution of systems with multiple and repeated right-hand sides. Preliminary performance results on a Linux Cluster suggest that the proposed Solution method is numerically scalable and demonstrate its potential for making the uncertainty quantification Of realistic systems tractable.

Why do birds participate in mixed-species foraging flocks? A large-scale synthesis

Mixed-species flocks of foraging birds have been documented from terrestrial habitats all over the world and are thought to form for either improved feeding efficiency or better protection from predators. Two kinds of flock participants are recognized: those that join other species ('followers') and are therefore likely to be the recipients of the benefits of flock participation and those that are joined ('leaders'). Through comparative analyses, using a large sample of flocks from around the world, we show that (1) 'followers' tend to be smaller, more insectivorous, and feed in higher strata than matched species that participate in flocks to a lesser extent and (2) 'leaders' tend to be cooperative breeders more often than matched species that are not known to lead flocks. Furthermore, meta-analyses of published results from across the world showed that bird species in terrestrial mixed-species flocks increase foraging rates and reduce vigilance compared to when they are solitary or in conspecific groups. Moreover, the increase in foraging rates is seen only with flock followers and not flock leaders. These findings suggest a role for predation in the evolution of mixed-species flocking. Species that are vulnerable to predation follow species whose vigilance they can exploit. By doing so, they are able to reduce their own vigilance and forage at higher rates. (C) 2009 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

PROBESYNC: Platform Based Synchronization for Enhanced Life of Large Scale Wireless Sensor Networks

Optimization in energy consumption of the existing synchronization mechanisms can lead to substantial gains in terms of network life in Wireless Sensor Networks (WSNs). In this paper, we analyze ERBS and TPSN, two existing synchronization algorithms for WSNs which use widely different approach, and compare their performance in large scale WSNs each of which consists of different type of platform and has varying node density. We, then, propose a novel algorithm, PROBESYNC, which takes advantage of differences in power required to transmit and receive a message on ERBS and TPSN and leverages the shortcomings of each of these algorithms. This leads to considerable improvement in energy conservation and enhanced life of large scale WSNs.

A novel approach for large-scale polypeptide folding based on elastic networks using continuous optimization

We present a new computationally efficient method for large-scale polypeptide folding using coarse-grained elastic networks and gradient-based continuous optimization techniques. The folding is governed by minimization of energy based on Miyazawa–Jernigan contact potentials. Using this method we are able to substantially reduce the computation time on ordinary desktop computers for simulation of polypeptide folding starting from a fully unfolded state. We compare our results with available native state structures from Protein Data Bank (PDB) for a few de-novo proteins and two natural proteins, Ubiquitin and Lysozyme. Based on our simulations we are able to draw the energy landscape for a small de-novo protein, Chignolin. We also use two well known protein structure prediction software, MODELLER and GROMACS to compare our results. In the end, we show how a modification of normal elastic network model can lead to higher accuracy and lower time required for simulation.

Large-scale structures in turbulent and reverse-transitional sink flow boundary layers

Aspects of large-scale organized structures in sink flow turbulent and reverse-transitional boundary layers are studied experimentally using hot-wire anemometry. Each of the present sink flow boundary layers is in a state of 'perfect equilibrium' or 'exact self-preservation' in the sense of Townsend (The Structure of Turbulent Shear Flow, 1st and 2nd edns, 1956, 1976, Cambridge University Press) and Rotta (Progr. Aeronaut. Sci., vol. 2, 1962, pp. 1-220) and conforms to the notion of 'pure wall-flow' (Coles, J. Aerosp. Sci., vol. 24, 1957, pp. 495-506), at least for the turbulent cases. It is found that the characteristic inclination angle of the structure undergoes a systematic decrease with the increase in strength of the streamwise favourable pressure gradient. Detectable wall-normal extent of the structure is found to be typically half of the boundary layer thickness. Streamwise extent of the structure shows marked increase as the favourable pressure gradient is made progressively severe. Proposals for the typical eddy forms in sink flow turbulent and reverse-transitional flows are presented, and the possibility of structural self-organization (i.e. individual hairpin vortices forming streamwise coherent hairpin packets) in these flows is also discussed. It is further indicated that these structural ideas may be used to explain, from a structural viewpoint, the phenomenon of soft relaminarization or reverse transition of turbulent boundary layers when subjected to strong streamwise favourable pressure gradients. Taylor's 'frozen turbulence' hypothesis is experimentally shown to be valid for flows in the present study even though large streamwise accelerations are involved, the flow being even reverse transitional in some cases. Possible conditions, which are required to be satisfied for the safe use of Taylor's hypothesis in pressure-gradient-driven flows, are also outlined. Measured convection velocities are found to be fairly close to the local mean velocities (typically 90% or more) suggesting that the structure gets convected downstream almost along with the mean flow.

PIV studies of large scale structures in the near field of small aspect ratio elliptic jets

The near flow field of small aspect ratio elliptic turbulent free jets (issuing from nozzle and orifice) was experimentally studied using a 2D PIV. Two point velocity correlations in these jets revealed the extent and orientation of the large scale structures in the major and minor planes. The spatial filtering of the instantaneous velocity field using Gaussian convolution kernel shows that while a single large vortex ring circumscribing the jet seems to be present at the exit of nozzle, the orifice jet exhibited a number of smaller vortex ring pairs close to jet exit. The smaller length scale observed in the case of the orifice jet is representative of the smaller azimuthal vortex rings that generate axial vortex field as they are convected. This results in the axis-switching in the case of orifice jet and may have a mechanism different from the self induction process as observed in the case of contoured nozzle jet flow.

Diagnosis of dominant forcing factors for large-scale vertical velocities during active and break phases of the monsoon

An attempt to diagnose the dominant forcings which drive the large-scale vertical velocities over the monsoon region has been made by computing the forcings like diabatic heating fields,etc. and the large-scale vertical velocities driven by these forcings for the contrasting periods of active and break monsoon situations; in order to understand the rainfall variability associated with them. Computation of diabatic heating fields show us that among different components of diabatic heating it is the convective heating that dominates at mid-tropospheric levels during an active monsoon period; whereas it is the sensible heating at the surface that is important during a break period. From vertical velocity calculations we infer that the prime differences in the large-scale vertical velocities seen throughout the depth of the atmosphere are due to the differences in the orders of convective heating; the maximum rate of latent heating being more than 10 degrees Kelvin per day during an active monsoon period; whereas during a break monsoon period it is of the order of 2 degrees Kelvin per day at mid-tropospheric levels. At low levels of the atmosphere, computations show that there is large-scale ascent occurring over a large spatial region, driven only by the dynamic forcing associated with vorticity and temperature advection during an active monsoon period. However, during a break monsoon period such large-scale spatial organization in rising motion is not seen. It is speculated that these differences in the low-level large-scale ascent might be causing differences in convective heating because the weaker the low level ascent, the lesser the convective instability which produces deep cumulus clouds and hence lesser the associated latent heat release. The forcings due to other components of diabatic heating, namely, the sensible heating and long wave radiative cooling do not influence the large-scale vertical velocities significantly.