ThunderRW

2021 ◽  
Vol 14 (11) ◽  
pp. 1992-2005 ◽  
Author(s):  
Shixuan Sun ◽  
Yuhang Chen ◽  
Shengliang Lu ◽  
Bingsheng He ◽  
Yuchen Li
Keyword(s):  
Random Walk ◽  
Programming Model ◽  
State Of The Art ◽  
Memory Access ◽  
The Core ◽  
Memory Efficiency ◽  
Order Of Magnitude ◽  
Graph Operation ◽  
Single Graph ◽  
Interleaving Technique

As random walk is a powerful tool in many graph processing, mining and learning applications, this paper proposes an efficient in-memory random walk engine named ThunderRW. Compared with existing parallel systems on improving the performance of a single graph operation, ThunderRW supports massive parallel random walks. The core design of ThunderRW is motivated by our profiling results: common RW algorithms have as high as 73.1% CPU pipeline slots stalled due to irregular memory access, which suffers significantly more memory stalls than the conventional graph workloads such as BFS and SSSP. To improve the memory efficiency, we first design a generic step-centric programming model named Gather-Move-Update to abstract different RW algorithms. Based on the programming model, we develop the step interleaving technique to hide memory access latency by switching the executions of different random walk queries. In our experiments, we use four representative RW algorithms including PPR, DeepWalk, Node2Vec and MetaPath to demonstrate the efficiency and programming flexibility of ThunderRW. Experimental results show that ThunderRW outperforms state-of-the-art approaches by an order of magnitude, and the step interleaving technique significantly reduces the CPU pipeline stall from 73.1% to 15.0%.

scholarly journals Small footprint optoelectrodes using ring resonators for passive light localization

2021 ◽  
Vol 7 (1) ◽  
Author(s):  
Vittorino Lanzio ◽  
Gregory Telian ◽  
Alexander Koshelev ◽  
Paolo Micheletti ◽  
Gianni Presti ◽  
...  
Keyword(s):  
State Of The Art ◽  
Three Dimensional ◽  
Network Activity ◽  
Ring Resonators ◽  
Spatiotemporal Resolution ◽  
Optical Ring Resonators ◽  
Light Localization ◽  
Order Of Magnitude ◽  
Small Footprint

AbstractThe combination of electrophysiology and optogenetics enables the exploration of how the brain operates down to a single neuron and its network activity. Neural probes are in vivo invasive devices that integrate sensors and stimulation sites to record and manipulate neuronal activity with high spatiotemporal resolution. State-of-the-art probes are limited by tradeoffs involving their lateral dimension, number of sensors, and ability to access independent stimulation sites. Here, we realize a highly scalable probe that features three-dimensional integration of small-footprint arrays of sensors and nanophotonic circuits to scale the density of sensors per cross-section by one order of magnitude with respect to state-of-the-art devices. For the first time, we overcome the spatial limit of the nanophotonic circuit by coupling only one waveguide to numerous optical ring resonators as passive nanophotonic switches. With this strategy, we achieve accurate on-demand light localization while avoiding spatially demanding bundles of waveguides and demonstrate the feasibility with a proof-of-concept device and its scalability towards high-resolution and low-damage neural optoelectrodes.

scholarly journals Representing Deep Neural Networks Latent Space Geometries with Graphs

Algorithms ◽  
2021 ◽  
Vol 14 (2) ◽  
pp. 39
Author(s):  
Carlos Lassance ◽  
Vincent Gripon ◽  
Antonio Ortega
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Deep Learning ◽  
Objective Function ◽  
Learning Process ◽  
Deep Neural Networks ◽  
State Of The Art ◽  
The Core ◽  
Learning Tasks ◽  
Latent Space

Deep Learning (DL) has attracted a lot of attention for its ability to reach state-of-the-art performance in many machine learning tasks. The core principle of DL methods consists of training composite architectures in an end-to-end fashion, where inputs are associated with outputs trained to optimize an objective function. Because of their compositional nature, DL architectures naturally exhibit several intermediate representations of the inputs, which belong to so-called latent spaces. When treated individually, these intermediate representations are most of the time unconstrained during the learning process, as it is unclear which properties should be favored. However, when processing a batch of inputs concurrently, the corresponding set of intermediate representations exhibit relations (what we call a geometry) on which desired properties can be sought. In this work, we show that it is possible to introduce constraints on these latent geometries to address various problems. In more detail, we propose to represent geometries by constructing similarity graphs from the intermediate representations obtained when processing a batch of inputs. By constraining these Latent Geometry Graphs (LGGs), we address the three following problems: (i) reproducing the behavior of a teacher architecture is achieved by mimicking its geometry, (ii) designing efficient embeddings for classification is achieved by targeting specific geometries, and (iii) robustness to deviations on inputs is achieved via enforcing smooth variation of geometry between consecutive latent spaces. Using standard vision benchmarks, we demonstrate the ability of the proposed geometry-based methods in solving the considered problems.

MapReduce Parallel Programming Model: A State-of-the-Art Survey

2015 ◽  
Vol 44 (4) ◽  
pp. 832-866 ◽  
Author(s):  
Ren Li ◽  
Haibo Hu ◽  
Heng Li ◽  
Yunsong Wu ◽  
Jianxi Yang
Keyword(s):  
Parallel Programming ◽  
Programming Model ◽  
State Of The Art ◽  
Parallel Programming Model

Two Orders of Magnitude Improvement in the Detection Limit of Droplet-Based Micro-Magnetofluidics with Planar Hall Effect Sensors

2021 ◽  
Vol 6 (1) ◽  
pp. 47
Author(s):  
Julian Schütt ◽  
Rico Illing ◽  
Oleksii Volkov ◽  
Tobias Kosub ◽  
Pablo Nicolás Granell ◽  
...  
Keyword(s):  
Detection Limit ◽  
Hall Effect ◽  
State Of The Art ◽  
Limit Of Detection ◽  
Research Field ◽  
High Throughput Analysis ◽  
Planar Hall Effect ◽  
Biologically Relevant ◽  
Sensing Platform ◽  
Order Of Magnitude

The detection, manipulation, and tracking of magnetic nanoparticles is of major importance in the fields of biology, biotechnology, and biomedical applications as labels as well as in drug delivery, (bio-)detection, and tissue engineering. In this regard, the trend goes towards improvements of existing state-of-the-art methodologies in the spirit of timesaving, high-throughput analysis at ultra-low volumes. Here, microfluidics offers vast advantages to address these requirements, as it deals with the control and manipulation of liquids in confined microchannels. This conjunction of microfluidics and magnetism, namely micro-magnetofluidics, is a dynamic research field, which requires novel sensor solutions to boost the detection limit of tiny quantities of magnetized objects. We present a sensing strategy relying on planar Hall effect (PHE) sensors in droplet-based micro-magnetofluidics for the detection of a multiphase liquid flow, i.e., superparamagnetic aqueous droplets in an oil carrier phase. The high resolution of the sensor allows the detection of nanoliter-sized superparamagnetic droplets with a concentration of 0.58 mg cm−3, even when they are only biased in a geomagnetic field. The limit of detection can be boosted another order of magnitude, reaching 0.04 mg cm−³ (1.4 million particles in a single 100 nL droplet) when a magnetic field of 5 mT is applied to bias the droplets. With this performance, our sensing platform outperforms the state-of-the-art solutions in droplet-based micro-magnetofluidics by a factor of 100. This allows us to detect ferrofluid droplets in clinically and biologically relevant concentrations, and even in lower concentrations, without the need of externally applied magnetic fields.

scholarly journals Comparison of the effectiveness of the Map-Reduce approach and the actor model in solving problems with high connectivity of inp data on the example of the optimization problem for a swarm of particles

2021 ◽  
pp. 049-055
Author(s):  
Larin V.O. ◽  
◽  
Provotar O.I. ◽  
Keyword(s):  
Optimization Problem ◽  
Map Reduce ◽  
Actor Model ◽  
Swarm Optimization ◽  
Memory Efficiency ◽  
Order Of Magnitude ◽  
High Connectivity ◽  
Memory Support ◽  
Bounded Input ◽  
Spark Framework

The paper defines the notion of distributed problems with bounded input components. Particle Swarm Optimization problem is shown to be an example of such a class. Such a problem's implementation based on the Map-Reduce model (implemented on the Spark framework) and an implementation based on an actor model with shared memory support (implemented on Strumok DSL) is provided. Both versions' performance assessment is conducted. The hybrid actor model is shown to be an order of magnitude more effective in time and memory efficiency than Map-Reduce implementation. Additional optimization for the hybrid actor model solution is proposed. The prospects of using the hybrid actor model for other similar problems are given

scholarly journals Optimization Learning: Perspective, Method, and Applications

2020 ◽  
Author(s):  
Risheng Liu
Keyword(s):  
Inverse Problems ◽  
Theoretical Analysis ◽  
Iterative Methods ◽  
State Of The Art ◽  
Learning Paradigm ◽  
Rigorous Analysis ◽  
The Core ◽  
Globally Convergent ◽  
Ill Posed ◽  
Art Performance

Numerous tasks at the core of statistics, learning, and vision areas are specific cases of ill-posed inverse problems. Recently, learning-based (e.g., deep) iterative methods have been empirically shown to be useful for these problems. Nevertheless, integrating learnable structures into iterations is still a laborious process, which can only be guided by intuitions or empirical insights. Moreover, there is a lack of rigorous analysis of the convergence behaviors of these reimplemented iterations, and thus the significance of such methods is a little bit vague. We move beyond these limits and propose a theoretically guaranteed optimization learning paradigm, a generic and provable paradigm for nonconvex inverse problems, and develop a series of convergent deep models. Our theoretical analysis reveals that the proposed optimization learning paradigm allows us to generate globally convergent trajectories for learning-based iterative methods. Thanks to the superiority of our framework, we achieve state-of-the-art performance on different real applications.

scholarly journals Asymmetric Distribution Measure for Few-shot Learning

2020 ◽  
Author(s):  
Wenbin Li ◽  
Lei Wang ◽  
Jing Huo ◽  
Yinghuan Shi ◽  
Yang Gao ◽  
...  
Keyword(s):  
State Of The Art ◽  
Asymmetric Distribution ◽  
Query Image ◽  
Local Descriptor ◽  
Innovative Design ◽  
Feature Representations ◽  
The Core ◽  
Measure Function ◽  
Asymmetric Relation ◽  
Core Idea

The core idea of metric-based few-shot image classification is to directly measure the relations between query images and support classes to learn transferable feature embeddings. Previous work mainly focuses on image-level feature representations, which actually cannot effectively estimate a class's distribution due to the scarcity of samples. Some recent work shows that local descriptor based representations can achieve richer representations than image-level based representations. However, such works are still based on a less effective instance-level metric, especially a symmetric metric, to measure the relation between a query image and a support class. Given the natural asymmetric relation between a query image and a support class, we argue that an asymmetric measure is more suitable for metric-based few-shot learning. To that end, we propose a novel Asymmetric Distribution Measure (ADM) network for few-shot learning by calculating a joint local and global asymmetric measure between two multivariate local distributions of a query and a class. Moreover, a task-aware Contrastive Measure Strategy (CMS) is proposed to further enhance the measure function. On popular miniImageNet and tieredImageNet, ADM can achieve the state-of-the-art results, validating our innovative design of asymmetric distribution measures for few-shot learning. The source code can be downloaded from https://github.com/WenbinLee/ADM.git.

scholarly journals Al Thumama Stadium: Local and Global Architectural Reach

2019 ◽  
Author(s):  
Hani Awni Hawamdeh
Keyword(s):  
State Of The Art ◽  
Vital Role ◽  
Local Identity ◽  
World Cup ◽  
The Core ◽  
Long Run ◽  
The World ◽  
Unique Experience ◽  
Core Mission

The world cup stadia have been a constant concern for the hosting countries. Many of them have become a burden on the economies of their countries, only to become white elephants after the tournaments end. Therefore, the core mission of the Supreme Committee for Delivery & Legacy in Qatar was to ensure that the World Cup Stadiums are built with a legacy and to remain functional in the long run, not just as facilities, but as cultural icons. Such efforts have promoted the exercise of stadia building in Qatar as a positive and unique experience. As a firm, we, at Arab Engineering Bureau, are honored to be part of the effort all through the making of Al Thumama Stadium, which will be discussed in this paper. Instead of a white elephant, Al Thumama Stadium is arguably a symbol of the local identity that will become part of the World Cup legacy, whilst being a state-of-the-art facility that plays a vital role in development of its surrounding neighborhood.

scholarly journals Trust in multi-agent systems

2004 ◽  
Vol 19 (1) ◽  
pp. 1-25 ◽  
Author(s):  
SARVAPALI D. RAMCHURN ◽  
DONG HUYNH ◽  
NICHOLAS R. JENNINGS
Keyword(s):  
Large Scale ◽  
State Of The Art ◽  
Comprehensive Treatment ◽  
Multi Agent Systems ◽  
Agent Based ◽  
Agent Systems ◽  
The Core ◽  
Open Distributed Systems ◽  
Multi Agent

Trust is a fundamental concern in large-scale open distributed systems. It lies at the core of all interactions between the entities that have to operate in such uncertain and constantly changing environments. Given this complexity, these components, and the ensuing system, are increasingly being conceptualised, designed, and built using agent-based techniques and, to this end, this paper examines the specific role of trust in multi-agent systems. In particular, we survey the state of the art and provide an account of the main directions along which research efforts are being focused. In so doing, we critically evaluate the relative strengths and weaknesses of the main models that have been proposed and show how, fundamentally, they all seek to minimise the uncertainty in interactions. Finally, we outline the areas that require further research in order to develop a comprehensive treatment of trust in complex computational settings.

New Instrumentation in Argonne's Hvem-Tandem Facility: Expanded Capability for in Situ Ion Beam Studies+

1995 ◽  
Vol 396 ◽  
Author(s):  
Charles W. Allen ◽  
Loren L. Funk ◽  
Edward A. Ryan
Keyword(s):  
Scientific Community ◽  
State Of The Art ◽  
Ion Irradiation ◽  
Ion Beam ◽  
Voltage Electron ◽  
User Facility ◽  
Order Of Magnitude ◽  
New Instrumentation ◽  
Better Than

AbstractDuring 1995, a state-of-the-art intermediate voltage electron microscope (IVEM) has been installed in the HVEM-Tandem Facility with in situ ion irradiation capabilities similar to those of the HVEM. A 300 kV Hitachi H-9000NAR has been interfaced to the two ion accelerators of the Facility, with a spatial resolution for imaging which is nearly an order of magnitude better than that for the 1.2 MV HVEM which dates from the early 1970s. The HVEM remains heavily utilized for electron- and ion irradiation-related materials studies, nevertheless, especially those for which less demanding microscopy is adequate. The capabilities and limitations of this IVEM and HVEM are compared. Both the HVEM and IVEM are part of the DOE funded User Facility and therefore are available to the scientific community for materials studies, free of charge for non-proprietary research.

Sign in / Sign up

Export Citation Format

Share Document