TURL

Predictive models of friction-induced vibration have proved elusive despite decades of research. There are many mechanisms that can cause brake squeal; friction coupled systems can be highly sensitive to small perturbations; and the dynamic properties of friction at the contact zone seem to be poorly understood. This paper describes experimental and theoretical work aimed at identifying the key ingredients of a predictive model. A large-scale experiment was carried out to identify squeal initiations using a pin-on-disc test rig: approximately 30,000 squeal initiations were recorded, covering a very wide range of frequencies. The theoretical model allows for completely general linear systems coupled at a single sliding point by friction: squeal is predicted using a linearised stability analysis. Results will be presented that show that almost all observed squeal events can be predicted within this model framework, but that some subsets require innovative friction modelling: predictions are highly dependent on the particular choice of friction model and its associated parameters.

Download Full-text

Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

10.36227/techrxiv.13025039.v1 ◽

2020 ◽

Author(s):

Yuan Yuan ◽

Lei Lin

Keyword(s):

Time Series ◽

Deep Learning ◽

Large Scale ◽

Temporal Structure ◽

Satellite Image ◽

Fine Tuning ◽

Small Scale ◽

Model Parameters ◽

Learning Approaches ◽

Wide Range

Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 1.91% to 6.69%. <div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

Download Full-text

The effect of cell tilting on turbulent thermal convection in a rectangular cell

Journal of Fluid Mechanics ◽

10.1017/jfm.2014.655 ◽

2014 ◽

Vol 762 ◽

pp. 273-287 ◽

Cited By ~ 15

Author(s):

Shuang-Xi Guo ◽

Sheng-Qi Zhou ◽

Xian-Rong Cen ◽

Ling Qu ◽

Yuan-Zheng Lu ◽

...

Keyword(s):

High Velocity ◽

Large Scale ◽

Velocity Measurements ◽

Cell Width ◽

Rectangular Cell ◽

Aspect Ratios ◽

Wide Range ◽

Almost All ◽

Temperature And Velocity Fields ◽

Pronounced Reduction

AbstractIn this study the influence of cell tilting on flow dynamics and heat transport is explored experimentally within a rectangular cell (aspect ratios ${\it\Gamma}_{x}=1$ and ${\it\Gamma}_{y}=0.25$). The measurements are carried out over a wide range of tilt angles ($0\leqslant {\it\beta}\leqslant {\rm\pi}/2\ \text{rad}$) at a constant Prandtl number ($\mathit{Pr}\simeq 6.3$) and Rayleigh number ($\mathit{Ra}\simeq 4.42\times 10^{9}$). The velocity measurements reveal that the large-scale circulation (LSC) is sensitive to the symmetry of the system. In the level case, the high-velocity band of the LSC concentrates at about a quarter of the cell width from the boundary. As the cell is slightly tilted (${\it\beta}\simeq 0.04\ \text{rad}$), the position of the high-velocity band quickly moves towards the boundary. With increasing ${\it\beta}$, the LSC changes gradually from oblique ellipse-like to square-like, and other more complicated patterns. Oscillations have been found in the temperature and velocity fields for almost all ${\it\beta}$, and are strongest at around ${\it\beta}\simeq 0.48\ \text{rad}$. As ${\it\beta}$ increases, the Reynolds number ($\mathit{Re}$) initially also increases, until it reaches its maximum at the transition angle ${\it\beta}=0.15\ \text{rad}$, after which it gradually decreases. The cell tilting causes a pronounced reduction of the Nusselt number ($\mathit{Nu}$). As ${\it\beta}$ increases from 0 to 0.15, 1.05 and ${\rm\pi}/2\ \text{rad}$, the reduction of $\mathit{Nu}$ is approximately 1.4 %, 5 % and 18 %, respectively. Over the ranges of $0\leqslant {\it\beta}\leqslant 0.15\ \text{rad}$, $0.15\leqslant {\it\beta}\leqslant 1.05\ \text{rad}$ and $1.05\leqslant {\it\beta}\leqslant {\rm\pi}/2\ \text{rad}$, the decay slopes are $8.57\times 10^{-2}$, $3.27\times 10^{-2}$ and $0.24\ \text{rad}^{-1}$, respectively.

Download Full-text

A general approach for improving deep learning-based medical relation extraction using a pre-trained model and fine-tuning

Database ◽

10.1093/database/baz116 ◽

2019 ◽

Vol 2019 ◽

Cited By ~ 2

Author(s):

Tao Chen ◽

Mingfen Wu ◽

Hexi Li

Keyword(s):

Deep Learning ◽

Large Scale ◽

Relation Extraction ◽

Training Model ◽

Biomedical Literature ◽

Training Data ◽

Fine Tuning ◽

Learning Approaches ◽

Additional Time ◽

Clinical Records

Abstract The automatic extraction of meaningful relations from biomedical literature or clinical records is crucial in various biomedical applications. Most of the current deep learning approaches for medical relation extraction require large-scale training data to prevent overfitting of the training model. We propose using a pre-trained model and a fine-tuning technique to improve these approaches without additional time-consuming human labeling. Firstly, we show the architecture of Bidirectional Encoder Representations from Transformers (BERT), an approach for pre-training a model on large-scale unstructured text. We then combine BERT with a one-dimensional convolutional neural network (1d-CNN) to fine-tune the pre-trained model for relation extraction. Extensive experiments on three datasets, namely the BioCreative V chemical disease relation corpus, traditional Chinese medicine literature corpus and i2b2 2012 temporal relation challenge corpus, show that the proposed approach achieves state-of-the-art results (giving a relative improvement of 22.2, 7.77, and 38.5% in F1 score, respectively, compared with a traditional 1d-CNN classifier). The source code is available at https://github.com/chentao1999/MedicalRelationExtraction.

Download Full-text

Design of HTS Magnetically Controlled Saturable Reactor

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.380-384.2982 ◽

2013 ◽

Vol 380-384 ◽

pp. 2982-2985

Author(s):

Hong Da Dong

Keyword(s):

High Temperature ◽

High Voltage ◽

Large Scale ◽

Reactive Power ◽

Magnetic Circuit ◽

Reactive Power Compensation ◽

Model Design ◽

Power Capacity ◽

Wide Range ◽

Small Capacity

There are many problems for traditional reactive power compensation devices to be applied in the grid, such as discontinuous adjustment, small capacity, complex control and harmonics. This paper aims to study a high temperature superconducting magnetically controlled saturable reactor (HTS MCSR), which has a wide range of stepless adjustment. It has a good application prospect in large scale reactive power compensation devices. Based on research of theory and core structure, a shaped-cylinder core is proposed. By means of calculation of saturable reactor and analysis of algebraic and magnetic circuit model, design of 220V HTS MCSR is finished. Results of normal conductive reactor prototype and simulations verify that the range of inductance adjustment is very wide. Furthermore, conceptual design of 35kV HTS MCSR confirms its reactive power capacity is so large, therefore, it is suitable for high voltage power system.

Download Full-text

Approximate weighted matching on emerging manycore and multithreaded architectures

The International Journal of High Performance Computing Applications ◽

10.1177/1094342012452893 ◽

2012 ◽

Vol 26 (4) ◽

pp. 413-430 ◽

Cited By ~ 20

Author(s):

Mahantesh Halappanavar ◽

John Feo ◽

Oreste Villa ◽

Antonino Tumeo ◽

Alex Pothen

Keyword(s):

Approximation Algorithms ◽

Large Scale ◽

Scientific Computing ◽

Graph Matching ◽

Combinatorial Problem ◽

Fine Tuning ◽

Matching Problem ◽

Weighted Matching ◽

Wide Range ◽

The Impact

Graph matching is a prototypical combinatorial problem with many applications in high-performance scientific computing. Optimal algorithms for computing matchings are challenging to parallelize. Approximation algorithms are amenable to parallelization and are therefore important to compute matchings for large-scale problems. Approximation algorithms also generate nearly optimal solutions that are sufficient for many applications. In this paper we present multithreaded algorithms for computing half-approximate weighted matching on state-of-the-art multicore (Intel Nehalem and AMD Magny-Cours), manycore (Nvidia Tesla and Nvidia Fermi), and massively multithreaded (Cray XMT) platforms. We provide two implementations: the first uses shared work queues and is suited for all platforms; and the second implementation, based on dataflow principles, exploits special features available on the Cray XMT. Using a carefully chosen dataset that exhibits characteristics from a wide range of applications, we show scalable performance across different platforms. In particular, for one instance of the input, an R-MAT graph (RMAT-G), we show speedups of about [Formula: see text] on [Formula: see text] cores of an AMD Magny-Cours, [Formula: see text] on [Formula: see text] cores of Intel Nehalem, [Formula: see text] on Nvidia Tesla and [Formula: see text] on Nvidia Fermi relative to one core of Intel Nehalem, and [Formula: see text] on [Formula: see text] processors of Cray XMT. We demonstrate strong as well as weak scaling for graphs with up to a billion edges using up to 12,800 threads. We avoid excessive fine-tuning for each platform and retain the basic structure of the algorithm uniformly across platforms. An exception is the dataflow algorithm designed specifically for the Cray XMT. To the best of the authors' knowledge, this is the first such large-scale study of the half-approximate weighted matching problem on multithreaded platforms. Driven by the critical enabling role of combinatorial algorithms such as matching in scientific computing and the emergence of informatics applications, there is a growing demand to support irregular computations on current and future computing platforms. In this context, we evaluate the capability of emerging multithreaded platforms to tolerate latency induced by irregular memory access patterns, and to support fine-grained parallelism via light-weight synchronization mechanisms. By contrasting the architectural features of these platforms against the Cray XMT, which is specifically designed to support irregular memory-intensive applications, we delineate the impact of these choices on performance.

Download Full-text

Performance Evaluation of Modularity Based Community Detection Algorithms in Large Scale Networks

Mathematical Problems in Engineering ◽

10.1155/2014/502809 ◽

2014 ◽

Vol 2014 ◽

pp. 1-15 ◽

Cited By ~ 2

Author(s):

Vinícius da Fonseca Vieira ◽

Carolina Ribeiro Xavier ◽

Nelson Francisco Favilla Ebecken ◽

Alexandre Gonçalves Evsukoff

Keyword(s):

Spectral Method ◽

Community Detection ◽

High Performance ◽

Large Scale ◽

Fine Tuning ◽

Detection Algorithms ◽

Research Areas ◽

Structure Detection ◽

Wide Range ◽

Large Scale Networks

Community structure detection is one of the major research areas of network science and it is particularly useful for large real networks applications. This work presents a deep study of the most discussed algorithms for community detection based on modularity measure: Newman’s spectral method using a fine-tuning stage and the method of Clauset, Newman, and Moore (CNM) with its variants. The computational complexity of the algorithms is analysed for the development of a high performance code to accelerate the execution of these algorithms without compromising the quality of the results, according to the modularity measure. The implemented code allows the generation of partitions with modularity values consistent with the literature and it overcomes 1 million nodes with Newman’s spectral method. The code was applied to a wide range of real networks and the performances of the algorithms are evaluated.

Download Full-text

Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

10.36227/techrxiv.13025039.v3 ◽

2020 ◽

Author(s):

Yuan Yuan ◽

Lei Lin

Keyword(s):

Time Series ◽

Deep Learning ◽

Large Scale ◽

Temporal Structure ◽

Satellite Image ◽

Fine Tuning ◽

Small Scale ◽

Model Parameters ◽

Learning Approaches ◽

Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

Download Full-text

Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

10.36227/techrxiv.13025039.v2 ◽

2020 ◽

Author(s):

Yuan Yuan ◽

Lei Lin

Keyword(s):

Time Series ◽

Deep Learning ◽

Large Scale ◽

Temporal Structure ◽

Satellite Image ◽

Fine Tuning ◽

Small Scale ◽

Model Parameters ◽

Learning Approaches ◽

Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

Download Full-text

Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

10.36227/techrxiv.13025039 ◽

2020 ◽

Author(s):

Yuan Yuan ◽

Lei Lin

Keyword(s):

Time Series ◽

Deep Learning ◽

Large Scale ◽

Temporal Structure ◽

Satellite Image ◽

Fine Tuning ◽

Small Scale ◽

Model Parameters ◽

Learning Approaches ◽

Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

Download Full-text