scholarly journals Cryptocurrency Price Estimation Using Hyperparameterized Oscillatory Activation Functions in LSTM Networks

2022 ◽  
Author(s):  
Pragya Mishra ◽  
Shubham Bharadwaj
Keyword(s):  
Performance Metrics ◽  
Computational Cost ◽  
Weighted Average ◽  
Gradient Flow ◽  
Activation Functions ◽  
Training Time ◽  
Price Estimation ◽  
Complex Decision ◽  
Critical Components ◽  
Decision Boundaries

Activation functions are critical components of neural networks, helping the model learn highly-intricate dependencies, trends, and patterns. Non-linear activation functions allow the model to behave as a functional approximator, learning complex decision boundaries and multi-dimensional patterns in the data. Activation functions can be combined with one another to learn better representations with the objective of improving gradient flow, performance metrics reducing training time and computational cost. Recent work on oscillatory activation functions\cite{noel2021growing}\cite{noel2021biologically} showcased their ability to perform competitively on image classification tasks using a compact architecture. Our work proposes the utilization of these oscillatory activation functions for predicting the volume-weighted average of Bitcoin on the G-Research Cryptocurrency Dataset. We utilize a popular LSTM architecture for this task achieving competitive results when compared to popular activation functions formally used.

scholarly journals Evaluation of Mixed Deep Neural Networks for Reverberant Speech Enhancement

Biomimetics ◽  
2019 ◽  
Vol 5 (1) ◽  
pp. 1 ◽  
Author(s):  
Michelle Gutiérrez-Muñoz ◽  
Astryd González-Salazar ◽  
Marvin Coto-Jiménez
Keyword(s):  
Neural Networks ◽  
Short Term Memory ◽  
Computational Cost ◽  
Real Life ◽  
Fixed Number ◽  
Training Procedure ◽  
Statistical Validation ◽  
Significant Drop ◽  
Training Time ◽  
Important Solution

Speech signals are degraded in real-life environments, as a product of background noise or other factors. The processing of such signals for voice recognition and voice analysis systems presents important challenges. One of the conditions that make adverse quality difficult to handle in those systems is reverberation, produced by sound wave reflections that travel from the source to the microphone in multiple directions. To enhance signals in such adverse conditions, several deep learning-based methods have been proposed and proven to be effective. Recently, recurrent neural networks, especially those with long short-term memory (LSTM), have presented surprising results in tasks related to time-dependent processing of signals, such as speech. One of the most challenging aspects of LSTM networks is the high computational cost of the training procedure, which has limited extended experimentation in several cases. In this work, we present a proposal to evaluate the hybrid models of neural networks to learn different reverberation conditions without any previous information. The results show that some combinations of LSTM and perceptron layers produce good results in comparison to those from pure LSTM networks, given a fixed number of layers. The evaluation was made based on quality measurements of the signal’s spectrum, the training time of the networks, and statistical validation of results. In total, 120 artificial neural networks of eight different types were trained and compared. The results help to affirm the fact that hybrid networks represent an important solution for speech signal enhancement, given that reduction in training time is on the order of 30%, in processes that can normally take several days or weeks, depending on the amount of data. The results also present advantages in efficiency, but without a significant drop in quality.

scholarly journals Binary and Multiclass Text Classification by Means of Separable Convolutional Neural Network

Inventions ◽  
2021 ◽  
Vol 6 (4) ◽  
pp. 70
Author(s):  
Elena Solovyeva ◽  
Ali Abdullah
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Convolutional Neural Network ◽  
Recurrent Neural Networks ◽  
Low Cost ◽  
Computational Cost ◽  
High Accuracy ◽  
Activation Functions ◽  
Fully Connected ◽  
Fully Connected Networks

In this paper, the structure of a separable convolutional neural network that consists of an embedding layer, separable convolutional layers, convolutional layer and global average pooling is represented for binary and multiclass text classifications. The advantage of the proposed structure is the absence of multiple fully connected layers, which is used to increase the classification accuracy but raises the computational cost. The combination of low-cost separable convolutional layers and a convolutional layer is proposed to gain high accuracy and, simultaneously, to reduce the complexity of neural classifiers. Advantages are demonstrated at binary and multiclass classifications of written texts by means of the proposed networks under the sigmoid and Softmax activation functions in convolutional layer. At binary and multiclass classifications, the accuracy obtained by separable convolutional neural networks is higher in comparison with some investigated types of recurrent neural networks and fully connected networks.

scholarly journals Evaluation of Mixed Deep Neural Networks for Reverberant Speech Enhancement

2019 ◽  
Author(s):  
Michelle Gutiérrez-Muñoz ◽  
Astryd González-Salazar ◽  
Marvin Coto-Jiménez
Keyword(s):  
Neural Networks ◽  
Short Term Memory ◽  
Computational Cost ◽  
Real Life ◽  
Fixed Number ◽  
Training Procedure ◽  
Statistical Validation ◽  
Training Time ◽  
Important Solution ◽  
Reverberant Speech

Speech signals are degraded in real-life environments, product of background noise or other factors. The processing of such signals for voice recognition and voice analysis systems presents important challenges. One of the conditions that make adverse quality difficult to handle in those systems is reverberation, produced by sound wave reflections that travel from the source to the microphone in multiple directions.To enhance signals in such adverse conditions, several deep learning-based methods have been proposed and proven to be effective. Recently, recurrent neural networks, especially those with long and short-term memory (LSTM), have presented surprising results in tasks related to time-dependent processing of signals, such as speech. One of the most challenging aspects of LSTM networks is the high computational cost of the training procedure, which has limited extended experimentation in several cases. In this work, we present a proposal to evaluate the hybrid models of neural networks to learn different reverberation conditions without any previous information. The results show that some combination of LSTM and perceptron layers produce good results in comparison to those from pure LSTM networks, given a fixed number of layers. The evaluation has been made based on quality measurements of the signal's spectrum, training time of the networks and statistical validation of results. Results help to affirm the fact that hybrid networks represent an important solution for speech signal enhancement, with advantages in efficiency, but without a significan drop in quality.

scholarly journals A consensus-based global optimization method for high dimensional machine learning problems

2020 ◽  
Author(s):  
Jose Carrillo ◽  
Shi Jin ◽  
Lei Li ◽  
Yuhua Zhu
Keyword(s):  
Optimization Problems ◽  
Computational Cost ◽  
Weighted Average ◽  
Planck Equation ◽  
Optimization Method ◽  
Learning Problems ◽  
High Dimensional ◽  
Nonconvex Functions ◽  
The Individual ◽  
Parameter Constraints

We improve recently introduced consensus-based optimization method, proposed in [R. Pinnau, C. Totzeck, O. Tse and S. Martin, Math. Models Methods Appl. Sci., 27(01):183{204, 2017], which is a gradient-free optimization method for general nonconvex functions. We rst replace the isotropic geometric Brownian motion by the component-wise one, thus removing the dimensionality dependence of the drift rate, making the method more competitive for high dimensional optimization problems. Secondly, we utilize the random mini-batch ideas to reduce the computational cost of calculating the weighted average which the individual particles tend to relax toward. For its mean- eld limit{a nonlinear Fokker-Planck equation{we prove, in both time continuous and semi-discrete settings, that the convergence of the method, which is exponential in time, is guaranteed with parameter constraints independent of the dimensionality. We also conduct numerical tests to high dimensional problems to check the success rate of the method.

scholarly journals Parametric space-frequency reduction for second-order system models

2018 ◽  
Author(s):  
Vanessa Cool ◽  
Frank Naets ◽  
Ward Rottiers ◽  
Wim Desmet
Keyword(s):  
Frequency Response ◽  
Cost Reduction ◽  
Frequency Response Function ◽  
Performance Metrics ◽  
Computational Cost ◽  
Order System ◽  
Model Parameters ◽  
Frequency Content ◽  
Space Frequency ◽  
Set Up

This research focusses on the computational cost reduction of frequency domain simulations in many-query applications with varying model parameters. These analyses are often encountered during the design of mechanical structures, where frequency response function (FRF) amplitudes are still one of the key performance metrics to be considered. Moreover, often inputs (number and frequency content) can vary broadly, which makes it all the more challenging to set up the reduced model.

scholarly journals Machine Learning Framework to Predict Chronic Kidney Disease using Ensemble Algorithm

2020 ◽  
Vol 9 (5) ◽  
pp. 1-6
Keyword(s):  
Machine Learning ◽  
Chronic Kidney Disease ◽  
Random Forest ◽  
Kidney Disease ◽  
Decision Tree ◽  
Performance Metrics ◽  
Weighted Average ◽  
Gradient Boosting ◽  
Support Vector ◽  
The Individual

Chronic Kidney Disease (CKD) is a worldwide concern that influences roughly 10% of the grown-up population on the world. For most of the people the early diagnosis of CKD is often not possible. Therefore, the utilization of present-day Computer aided supported strategies is important to help the conventional CKD finding framework to be progressively effective and precise. In this project, six modern machine learning techniques namely Multilayer Perceptron Neural Network, Support Vector Machine, Naïve Bayes, K-Nearest Neighbor, Decision Tree, Logistic regression were used and then to enhance the performance of the model Ensemble Algorithms such as ADABoost, Gradient Boosting, Random Forest, Majority Voting, Bagging and Weighted Average were used on the Chronic Kidney Disease dataset from the UCI Repository. The model was tuned finely to get the best hyper parameters to train the model. The performance metrics used to evaluate the model was measured using Accuracy, Precision, Recall, F1-score, Mathew`s Correlation Coefficient and ROC-AUC curve. The experiment was first performed on the individual classifiers and then on the Ensemble classifiers. The ensemble classifier like Random Forest and ADABoost performed better with 100% Accuracy, Precision and Recall when compared to the individual classifiers with 99.16% accuracy, 98.8% Precision and 100% Recall obtained from Decision Tree Algorithm

scholarly journals Communication-Efficient Stochastic Gradient MCMC for Neural Networks

2019 ◽  
Vol 33 ◽  
pp. 4173-4180 ◽  
Author(s):  
Chunyuan Li ◽  
Changyou Chen ◽  
Yunchen Pu ◽  
Ricardo Henao ◽  
Lawrence Carin
Keyword(s):  
Neural Networks ◽  
Probability Distributions ◽  
Computational Cost ◽  
Time Estimation ◽  
Stochastic Gradient ◽  
Communication Overhead ◽  
Test Accuracy ◽  
Training Time ◽  
Learning Probability ◽  
Policy Optimization

Learning probability distributions on the weights of neural networks has recently proven beneficial in many applications. Bayesian methods such as Stochastic Gradient Markov Chain Monte Carlo (SG-MCMC) offer an elegant framework to reason about model uncertainty in neural networks. However, these advantages usually come with a high computational cost. We propose accelerating SG-MCMC under the masterworker framework: workers asynchronously and in parallel share responsibility for gradient computations, while the master collects the final samples. To reduce communication overhead, two protocols (downpour and elastic) are developed to allow periodic interaction between the master and workers. We provide a theoretical analysis on the finite-time estimation consistency of posterior expectations, and establish connections to sample thinning. Our experiments on various neural networks demonstrate that the proposed algorithms can greatly reduce training time while achieving comparable (or better) test accuracy/log-likelihood levels, relative to traditional SG-MCMC. When applied to reinforcement learning, it naturally provides exploration for asynchronous policy optimization, with encouraging performance improvement.

scholarly journals Malware Classification Based on Shallow Neural Network

2020 ◽  
Vol 12 (12) ◽  
pp. 219
Author(s):  
Pin Yang ◽  
Huiyu Zhou ◽  
Yue Zhu ◽  
Liang Liu ◽  
Lei Zhang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Computational Cost ◽  
Malicious Code ◽  
Classification Model ◽  
Evolutionary Trend ◽  
Training Time ◽  
Binary File ◽  
Malware Classification ◽  
N Gram

The emergence of a large number of new malicious code poses a serious threat to network security, and most of them are derivative versions of existing malicious code. The classification of malicious code is helpful to analyze the evolutionary trend of malicious code families and trace the source of cybercrime. The existing methods of malware classification emphasize the depth of the neural network, which has the problems of a long training time and large computational cost. In this work, we propose the shallow neural network-based malware classifier (SNNMAC), a malware classification model based on shallow neural networks and static analysis. Our approach bridges the gap between precise but slow methods and fast but less precise methods in existing works. For each sample, we first generate n-grams from their opcode sequences of the binary file with a decompiler. An improved n-gram algorithm based on control transfer instructions is designed to reduce the n-gram dataset. Then, the SNNMAC exploits a shallow neural network, replacing the full connection layer and softmax with the average pooling layer and hierarchical softmax, to learn from the dataset and perform classification. We perform experiments on the Microsoft malware dataset. The evaluation result shows that the SNNMAC outperforms most of the related works with 99.21% classification precision and reduces the training time by more than half when compared with the methods using DNN (Deep Neural Networks).

Handling Large Medical Data Sets for Disease Detection

2011 ◽  
pp. 162-176 ◽  
Author(s):  
Rahul Kala ◽  
Anupam Shukla ◽  
Ritu Tiwari
Keyword(s):  
Breast Cancer ◽  
Machine Learning ◽  
Biomedical Engineering ◽  
Intelligent Systems ◽  
Computational Cost ◽  
Poor Performance ◽  
Medical Data ◽  
Data Sets ◽  
Training Time ◽  
Low Computational Cost

The breakthrough in the field of intelligent systems has spread its fruits to the field of biomedical engineering as well; where a series of models are being applied to automatically detect diseases based on some parameters or inputs. The continuous research in this field has resulted in a large amount of database being created for many diseases which becomes very difficult to train. Also the number of attributes is under constant rise. This increases the dimensionality of the problem and ultimately leads to poor performance. In this chapter we deal with the methods to handle these situations. We discuss the mechanism to divide data between different sub-systems. We also discuss the method of division of the attributes to reduce the training time and complexity. The resultant systems are able to train better due to low computational cost and hence give better performance. We validated this with the Breast Cancer database from the UCI Machine Learning repository and found our algorithm optimal.

A CNN With Deep Learning for Non-Equilibrium Characterization of Al-Sm Melt Infusion Into a B4C Packed Bed

2021 ◽  
Author(s):  
Julio Aguilar ◽  
Laura Sandoval ◽  
Arturo Rodriguez ◽  
Sanjay Shantha Kumar ◽  
Jose Terrazas ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
High Temperature ◽  
Convolutional Neural Network ◽  
Computational Cost ◽  
Packed Bed ◽  
Reaction Products ◽  
Training Time ◽  
Ti Alloys ◽  
Prediction Approach

Abstract In seeking predictability of characterizing materials for ultra-high temperature materials for hypersonic vehicles, the use of the convolutional neural network for characterizing the behavior of liquid Al-Sm-X (Hf, Zr, Ti) alloys within a B4C packed to determine the reaction products for which they are usually done with the scanning electron microscope (SEM) or X-ray diffraction (XRD) at ultra-high temperatures (> 1600°C). Our goal is to predict ultimately the products as liquid Al-Sm-X (Hf, Zr, Ti) alloys infiltrate into a B4C packed bed. Material characterization determines the processing path and final species from the reacting infusion consisting of fluid flow through porous channels, consumption of elemental components, and reaction forming boride and carbide precipitates. Since characterization is time-consuming, an expert in this field is required; our approach is to characterize and track these species using a Convolutional Neural Network (CNN) to facilitate and automate analysis of images. Although Deep Learning seems to provide an automated prediction approach, some of these challenges faced under this research are difficult to overcome. These challenges include data required, accuracy, training time, and computational cost requirements for a CNN. Our approach was to perform experiments on high-temperature metal infusion under B4C Packed Bed infiltration in a parametric matrix of cases. We characterized images using SEM and XRD images and run/optimize our CNN, which yields an innovative method for characterization via Deep Learning compared to traditional practices.

Sign in / Sign up

Export Citation Format

Share Document