scholarly journals Detection of Anomalous Diffusion with Deep Residual Networks

Entropy ◽  
2021 ◽  
Vol 23 (6) ◽  
pp. 649
Author(s):  
Miłosz Gajowczyk ◽  
Janusz Szwabiński
Keyword(s):  
Image Classification ◽  
Anomalous Diffusion ◽  
Numerical Experiments ◽  
Driving Forces ◽  
Living Cells ◽  
Diffusion Type ◽  
Training Time ◽  
Initial Network ◽  
Unseen Data ◽  
Insight Into

Identification of the diffusion type of molecules in living cells is crucial to deduct their driving forces and hence to get insight into the characteristics of the cells. In this paper, deep residual networks have been used to classify the trajectories of molecules. We started from the well known ResNet architecture, developed for image classification, and carried out a series of numerical experiments to adapt it to detection of diffusion modes. We managed to find a model that has a better accuracy than the initial network, but contains only a small fraction of its parameters. The reduced size significantly shortened the training time of the model. Moreover, the resulting network has less tendency to overfitting and generalizes better to unseen data.

Signal processing algorithm for neural networks with integrodifferential splines as an activation function and its particular case of image classification

2021 ◽  
Author(s):  
T.K. Biryukova
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Image Classification ◽  
Activation Function ◽  
Experimental Comparison ◽  
Training Time ◽  
Operation Speed ◽  
The Neural Network ◽  
Linear Algebraic Equations ◽  
Network Operation

Classic neural networks suppose trainable parameters to include just weights of neurons. This paper proposes parabolic integrodifferential splines (ID-splines), developed by author, as a new kind of activation function (AF) for neural networks, where ID-splines coefficients are also trainable parameters. Parameters of ID-spline AF together with weights of neurons are vary during the training in order to minimize the loss function thus reducing the training time and increasing the operation speed of the neural network. The newly developed algorithm enables software implementation of the ID-spline AF as a tool for neural networks construction, training and operation. It is proposed to use the same ID-spline AF for neurons in the same layer, but different for different layers. In this case, the parameters of the ID-spline AF for a particular layer change during the training process independently of the activation functions (AFs) of other network layers. In order to comply with the continuity condition for the derivative of the parabolic ID-spline on the interval (x x0, n) , its parameters fi (i= 0,...,n) should be calculated using the tridiagonal system of linear algebraic equations: To solve the system it is necessary to use two more equations arising from the boundary conditions for specific problems. For exam- ple the values of the grid function (if they are known) in the points (x x0, n) may be used for solving the system above: f f x0 = ( 0) , f f xn = ( n) . The parameters Iii+1 (i= 0,...,n−1 ) are used as trainable parameters of neural networks. The grid boundaries and spacing of the nodes of ID-spline AF are best chosen experimentally. The optimal selection of grid nodes allows improving the quality of results produced by the neural network. The formula for a parabolic ID-spline is such that the complexity of the calculations does not depend on whether the grid of nodes is uniform or non-uniform. An experimental comparison of the results of image classification from the popular FashionMNIST dataset by convolutional neural 0, x< 0 networks with the ID-spline AFs and the well-known ReLUx( ) =AF was carried out. The results reveal that the usage x x, ≥ 0 of the ID-spline AFs provides better accuracy of neural network operation than the ReLU AF. The training time for two convolutional layers network with two ID-spline AFs is just about 2 times longer than with two instances of ReLU AF. Doubling of the training time due to complexity of the ID-spline formula is the acceptable price for significantly better accuracy of the network. Wherein the difference of an operation speed of the networks with ID-spline and ReLU AFs will be negligible. The use of trainable ID-spline AFs makes it possible to simplify the architecture of neural networks without losing their efficiency. The modification of the well-known neural networks (ResNet etc.) by replacing traditional AFs with ID-spline AFs is a promising approach to increase the neural network operation accuracy. In a majority of cases, such a substitution does not require to train the network from scratch because it allows to use pre-trained on large datasets neuron weights supplied by standard software libraries for neural network construction thus substantially shortening training time.

Sustainability Reporting Driving Forces and Management Control Systems

2016 ◽  
Vol 28 (2) ◽  
pp. 103-124 ◽  
Author(s):  
Irene M. Herremans ◽  
Jamal A. Nazari
Keyword(s):  
Control Systems ◽  
Oil And Gas ◽  
Driving Forces ◽  
Sustainability Reporting ◽  
Oil And Gas Industry ◽  
Management Control ◽  
Gas Industry ◽  
Stakeholder Relationships ◽  
External Pressures ◽  
Insight Into

ABSTRACT This study investigates how seemingly similar external pressures elicited diverse sustainability reporting control systems and processes in a sample of Canadian companies in the oil and gas industry. Using interviews with companies and their stakeholders, we found that the type of sustainability reporting control systems depended on the managerial motivations and attitudes within companies as they responded to external pressures. More specifically, our results provide insight into how formal and informal sustainability reporting control systems were developed according to various managerial motivations and different types of stakeholder relationships. The type and balance between formal and informal control systems, in turn, influenced the sustainability reporting characteristics that the company was able to develop. We contribute to the literature by differentiating companies based on their institutional logics to deal with external pressures, managerial motivations, and stakeholder relationships, that in turn influenced their control system characteristics including reporting structures, information systems, and assurances.

Improve Language Modeling for Code Completion Through Learning General Token Repetition of Source Code with Optimized Memory

2019 ◽  
Vol 29 (11n12) ◽  
pp. 1801-1818
Author(s):  
Yixiao Yang ◽  
Xiang Chen ◽  
Jiaguang Sun
Keyword(s):  
Prediction Accuracy ◽  
Language Model ◽  
Source Code ◽  
Attention Mechanism ◽  
Data Set ◽  
Training Time ◽  
Unseen Data ◽  
Code Completion ◽  
High Chance ◽  
Better Than

In last few years, applying language model to source code is the state-of-the-art method for solving the problem of code completion. However, compared with natural language, code has more obvious repetition characteristics. For example, a variable can be used many times in the following code. Variables in source code have a high chance to be repetitive. Cloned code and templates, also have the property of token repetition. Capturing the token repetition of source code is important. In different projects, variables or types are usually named differently. This means that a model trained in a finite data set will encounter a lot of unseen variables or types in another data set. How to model the semantics of the unseen data and how to predict the unseen data based on the patterns of token repetition are two challenges in code completion. Hence, in this paper, token repetition is modelled as a graph, we propose a novel REP model which is based on deep neural graph network to learn the code toke repetition. The REP model is to identify the edge connections of a graph to recognize the token repetition. For predicting the token repetition of token [Formula: see text], the information of all the previous tokens needs to be considered. We use memory neural network (MNN) to model the semantics of each distinct token to make the framework of REP model more targeted. The experiments indicate that the REP model performs better than LSTM model. Compared with Attention-Pointer network, we also discover that the attention mechanism does not work in all situations. The proposed REP model could achieve similar or slightly better prediction accuracy compared to Attention-Pointer network and consume less training time. We also find other attention mechanism which could further improve the prediction accuracy.

The origin of different driving forces between O–H/N–H functional groups in metal ligand cooperation: mechanistic insight into Mn(i) catalysed transfer hydrogenation

2020 ◽  
Vol 10 (1) ◽  
pp. 169-179 ◽  
Author(s):  
Li Zhou ◽  
Datai Liu ◽  
Haiyi Lan ◽  
Xiujian Wang ◽  
Cunyuan Zhao ◽  
...  
Keyword(s):  
Catalytic Activity ◽  
Functional Groups ◽  
Driving Forces ◽  
Transfer Hydrogenation ◽  
Bifunctional Catalysts ◽  
Mechanistic Insight ◽  
Insight Into ◽  
Metal Ligand

The origin of different catalytic activity between two structurally similar Lewis basic bifunctional catalysts.

scholarly journals Nucleosome–nucleosome interactions via histone tails and linker DNA regulate nuclear rigidity

2017 ◽  
Vol 28 (11) ◽  
pp. 1580-1589 ◽  
Author(s):  
Yuta Shimamoto ◽  
Sachiko Tamura ◽  
Hiroshi Masumoto ◽  
Kazuhiro Maeshima
Keyword(s):  
Molecular Mechanisms ◽  
Mechanical Response ◽  
Nuclear Architecture ◽  
Living Cells ◽  
Genome Integrity ◽  
Biological Processes ◽  
Cell Nuclei ◽  
Histone Tails ◽  
Insight Into

Cells, as well as the nuclei inside them, experience significant mechanical stress in diverse biological processes, including contraction, migration, and adhesion. The structural stability of nuclei must therefore be maintained in order to protect genome integrity. Despite extensive knowledge on nuclear architecture and components, however, the underlying physical and molecular mechanisms remain largely unknown. We address this by subjecting isolated human cell nuclei to microneedle-based quantitative micromanipulation with a series of biochemical perturbations of the chromatin. We find that the mechanical rigidity of nuclei depends on the continuity of the nucleosomal fiber and interactions between nucleosomes. Disrupting these chromatin features by varying cation concentration, acetylating histone tails, or digesting linker DNA results in loss of nuclear rigidity. In contrast, the levels of key chromatin assembly factors, including cohesin, condensin II, and CTCF, and a major nuclear envelope protein, lamin, are unaffected. Together with in situ evidence using living cells and a simple mechanical model, our findings reveal a chromatin-based regulation of the nuclear mechanical response and provide insight into the significance of local and global chromatin structures, such as those associated with interdigitated or melted nucleosomal fibers.

Hyperspectral Image Classification Based on Artificial Immune System

2012 ◽  
Vol 500 ◽  
pp. 806-812 ◽  
Author(s):  
Farhad Samadzadegan ◽  
Shahin Rahmatollahi Namin ◽  
Mohammad Ali Rajabi
Keyword(s):  
Immune System ◽  
Image Classification ◽  
Supervised Classification ◽  
Hyperspectral Image ◽  
Artificial Immune Systems ◽  
Artificial Immune ◽  
Hyperspectral Image Classification ◽  
Spectral Dimensionality ◽  
Hyperspectral Remote Sensing Images ◽  
Insight Into

The high spectral dimensionality in hyperspectral images causes the reduction of accuracy for common statistical classification methods in these images. Hence the generation and implementation of more complicated methods have gained great importance in this field. One of these methods is the Artificial Immune Systems which is inspired by natural immune system. Despite its great potentiality, it is rarely utilized for spatial sciences and image classification. In this paper a supervised classification algorithm with the application of hyperspectral remote sensing images is proposed. In order to gain better insight into its capability, its accuracy is compared with Artificial Neural Network. The results show better image classification accuracy for the Artificial Immune method.

THE INTERACTION POTENTIAL BETWEEN MICRO/NANO CURVED SURFACE BODY WITH NEGATIVE GAUSS CURVATURE AND AN OUTSIDE PARTICLE

2015 ◽  
Vol 15 (06) ◽  
pp. 1540055
Author(s):  
DAN WANG ◽  
YAJUN YIN ◽  
JIYE WU ◽  
ZHENG ZHONG
Keyword(s):  
Interaction Potential ◽  
Numerical Experiments ◽  
Driving Forces ◽  
Pair Potential ◽  
Gauss Curvature ◽  
Curved Surface ◽  
Negative Exponential

Based on the negative exponential pair potential ([Formula: see text]), the interaction potential between curved surface body with negative Gauss curvature and an outside particle is proved to be of curvature-based form, i.e., it can be written as a function of curvatures. Idealized numerical experiments are designed to test the accuracy of the curvature-based potential. Compared with the previous results, it is confirmed that the interaction potential between curved surface body and an outside particle has a unified expression of curvatures regardless of the sign of Gauss curvature. Further, propositions below are confirmed: Highly curved surface body may induce driving forces, curvatures and the gradient of curvatures are the essential factors forming the driving forces.

scholarly journals A Synthetic Microbial Operational Amplifier

2017 ◽  
Author(s):  
Ji Zeng ◽  
Jaewook Kim ◽  
Areen Banerjee ◽  
Rahul Sarpeshkar
Keyword(s):  
Transcription Factor ◽  
Synthetic Biology ◽  
Operational Amplifier ◽  
Biological Systems ◽  
Logic Gates ◽  
Living Cells ◽  
Output Stage ◽  
Cellular Circuits ◽  
Differential Transcription ◽  
Insight Into

AbstractSynthetic biology has created oscillators, latches, logic gates, logarithmically linear circuits, and load drivers that have electronic analogs in living cells. The ubiquitous operational amplifier, which allows circuits to operate robustly and precisely has not been built with bio-molecular parts. As in electronics, a biological operational-amplifier could greatly improve the predictability of circuits despite noise and variability, a problem that all cellular circuits face. Here, we show how to create a synthetic 3-stage inducer-input operational amplifier with a differential transcription-factor stage, a CRISPR-based push-pull stage, and an enzymatic output stage with just 5 proteins including dCas9. Our ‘Bio-OpAmp’ expands the toolkit of fundamental circuits available to bioengineers or biologists, and may shed insight into biological systems that require robust and precise molecular homeostasis and regulation.One Sentence SummaryA synthetic bio-molecular operational amplifier that can enable robust, precise, and programmable homeostasis and regulation in living cells with just 5 protein parts is described.

scholarly journals Decolonising Arts and Culture in Belgium

2021 ◽  
Vol 30 (2) ◽  
pp. 132-140
Author(s):  
Axel Mudahemuka Gossiaux
Keyword(s):  
Information Technology ◽  
Social Media ◽  
Driving Forces ◽  
Black Music ◽  
Arts And Culture ◽  
Contextual Elements ◽  
The City ◽  
Insight Into ◽  
Do So

This contribution gives insight into the decolonisation of thought by presenting Black Out, a transmedia initiative located in the city of Liège in Belgium. Black Out is a project designed for promoting black music and culture and fighting against racism, principally through information technology and social media. I highlight how Black Out may participate in efforts for decolonising arts and culture in Belgium and Europe. To do so, I present a few contextual elements about racism and the postcolonial debate in Belgium before giving examples on how the projects of Black Out are in line with some of the driving forces of the decolonial approach.

Sign in / Sign up

Export Citation Format

Share Document