NESTED SPARSE BAYESIAN LEARNING FOR BLOCK-SPARSE SIGNALS WITH INTRA-BLOCK CORRELATION

In this work, we address the recovery of block sparse vectors with intra-block correlation, i.e., the recovery of vectors in which the correlated nonzero entries are constrained to lie in a few clusters, from noisy underdetermined linear measurements. Among Bayesian sparse recovery techniques, the cluster Sparse Bayesian Learning (SBL) is an efficient tool for block-sparse vector recovery, with intra-block correlation. However, this technique uses a heuristic method to estimate the intra-block correlation. In this paper, we propose the Nested SBL (NSBL) algorithm, which we derive using a novel Bayesian formulation that facilitates the use of the monotonically convergent nested Expectation Maximization (EM) and a Kalman filtering based learning framework. Unlike the cluster-SBL algorithm, this formulation leads to closed-form EMupdates for estimating the correlation coefficient. We demonstrate the efficacy of the proposed NSBL algorithm using Monte Carlo simulations.

Layered Exact-Repair Regenerating Codes via Embedded Error Correction and Block Designs

A new class of exact-repair regenerating codes is constructed by stitching together shorter erasure correction codes, where the stitching pattern can be viewed as block designs. The proposed codes have the help-by-transfer property where the helper nodes simply transfer part of the stored data directly, without performing any computation. This embedded error correction structure makes the decoding process straightforward, and in some cases the complexity is very low. We show that this construction is able to achieve performance better than space-sharing between the minimum storage regenerating codes and the minimum repair-bandwidth regenerating codes, and it is the first class of codes to achieve this performance. In fact, it is shown that the proposed construction can achieve a nontrivial point on the optimal functional-repair tradeoff, and it is asymptotically optimal at high rate, i.e., it asymptotically approaches the minimum storage and the minimum repair-bandwidth simultaneously.

Adaptive executions of hyperbolic block-structured AMR applications on GPU systems

A block-structured adaptive mesh refinement (AMR) technique has been used to obtain numerical solutions for many scientific applications. Some block-structured AMR approaches have focused on forming patches of non-uniform sizes where the size of a patch can be tuned to the geometry of a region of interest. In this paper, we develop strategies for adaptive execution of block-structured AMR applications on GPUs, for hyperbolic directionally split solvers. While effective hybrid execution strategies exist for applications with uniform patches, our work considers efficient execution of non-uniform patches with different workloads. Our techniques include bin-packing work units to load balance GPU computations, adaptive asynchronism between CPU and GPU executions using a knapsack formulation, and scheduling communications for multi-GPU executions. Our experiments with synthetic and real data, for single-GPU and multi-GPU executions, on Tesla S1070 and Fermi C2070 clusters, show that our strategies result in up to a 3.23 speedup in performance over existing strategies.

Seismic response of reduced-scale modular block and rigid faced reinforced walls through shaking table tests

This paper focuses on understanding the seismic response of geosynthetic reinforced retaining walls through shaking table tests on models of modular block and rigid faced reinforced retaining walls. Reduced-scale models of retaining walls reinforced with geogrid layers were constructed in a laminar box mounted on a uniaxial shaking table and subjected to various levels of sinusoidal base shaking. Models were instrumented with ultrasonic displacement sensors, earth pressure sensors and accelerometers. Effects of backfill density, number of reinforcement layers and reinforcement type on the performance of rigid faced and modular block walls were studied through different series of model tests. Performances of the walls were assessed in terms of face deformations, crest settlement and acceleration amplification at different elevations and compared. Modular block walls performed better than the rigid faced walls for the same level of base shaking because of the additional support derived by stacking the blocks with an offset. Type and quantity of reinforcement has significant effect on the seismic performance of both the types of walls. Displacements are more sensitive to relative density of the backfill and decrease with increasing relative density, the effect being more pronounced in case of unreinforced walls compared to the reinforced ones. Acceleration amplifications are not affected by the wall facing and inclusion of reinforcement. (C) 2015 Elsevier Ltd. All rights reserved.

Neoarchean arc magmatism followed by high-temperature, high-pressure metamorphism in the Nilgiri Block, southern India

The Nilgiri Block, southern India is an exhumed lower crust formed through arc magmatic processes in the Neoarchean. The main lithologies in this terrane include charnockites, gneisses, volcanic tuff, metasediments, banded iron formation and mafic-ultramafic bodies. Mafic-ultramafic rocks are present towards the northern and central part of the Nilgiri Block. We examine the evolution of these mafic granulites/metagabbros by phase diagram modeling and U-Pb sensitive high resolution ion microprobe (SHRIMP) dating. They consist of a garnet-clinopyroxene-plagioclase-hornblende-ilmenite +/- orthopyroxene +/- rutile assemblage. Garnet and clinopyroxene form major constituents with labradorite and orthopyroxene as the main mineral inclusions. Labradorite, identified using Raman analysis, shows typical peaks at 508 cm(-1), 479 cm(-1), 287 cm(-1) and 177 cm(-1). It is stable along with orthopyroxene towards the low-pressure high-temperature region of the granulite fades (M1 stage). Subsequently, orthopyroxene reacted with plagioclase to form the peak garnet + clinopyroxene + rutile assemblage (M2 stage). The final stage is represented by amphibolite facies-hornblende and plagioclase-rim around the garnet-clinopyroxene assemblage (M3 stage). Phase diagram modeling shows that these mafic granulites followed an anticlockwise P-T-t path during their evolution. The initial high-temperature metamorphism (M1 stage) was at 850-900 degrees C and similar to 9 kbar followed by high-pressure granulite fades metamorphism (M2 stage) at 850-900 degrees C and 14-15 kbar. U-Pb isotope studies of zircons using SHRIMP revealed late Neoarchean to early paleoproterozoic ages of crystallization and metamorphism respectively. The age data shows that these mafic granulites have undergone arc magmatism at ca. 25392 +/- 3 Ma and high-temperature, high-pressure metamorphism at ca. 2458.9 +/- 8.6 Ma. Thus our results suggests a late Neoarchean arc magmatism followed by early paleoproterozoic high-temperature, high-pressure granulite facies metamorphism due to the crustal thickening and suturing of the Nilgiri Block onto the Dharwar Craton. (C) 2015 Elsevier B.V. All rights reserved.

Archean tectonics and crustal evolution of the Biligiri Rangan Block, southern India

The Southern Granulite Terrain in India is a collage of crustal blocks ranging in age from Archean to Neoproterozoic. This study investigate the tectonic evolution of one of the northernmost block- the Biligiri Block (BRB) through a multidisciplinary approach involving field investigation, petrographic studies, LA-ICPMS zircon U-Pb geochronology, Hf isotopic analyses, metamorphic P-T phase diagram computations, and crustal thickness modeling. The garnet bearing quartzofeldspathic gneiss from the central BRB preserve Mesoarchean magmatic zircons with ages between 3207 and 2806 Ma and positive epsilon Hf value (+2.7) which possibly indicates vestiges of a Mesoarchean primitive continental crust. The occurrence of quartzite-iron formation intercalation as well as ultramafic lenses along the western boundary of the BRB is interpreted to indicate that the Kollegal structural lineament is a possible paleo-suture. Phase diagram computation of a metagabbro from the southwestern periphery of the Kollegal suture zone reveals high-pressure (similar to 18.5 kbar) and medium-temperature (similar to 840 degrees C) metamorphism, likely during eastward subduction of the Western Dharwar oceanic crust beneath the Mesoarchean BRB. In the model presented here, slab subduction, melting and underplating processes generated arc magmatism and subsequent charnockitization within the BRB between ca. 2650 Ma and ca. 2498 Ma. These results thus reveal Meso- to Neoarchean tectonic evolution of the BRB. The spatial variation of crustal thickness, derived from flexure inversion technique, provides additional constraints on the tectonic linkage of the BRB with its surrounding terrains. In conjunction with published data, the Moyar and the Kollegal suture zones are considered to mark the trace of ocean closure along which the Nilgiri and Biligiri Rangan Blocks accreted on to the Western Dharwar Craton. (C) 2016 Elsevier B.V. All rights reserved.