scholarly journals Estudo de algoritmos de otimização inspirados na natureza aplicados ao treinamento de redes neurais artificiais

2021 ◽  
Author(s):  
◽  
M. F. Bouzon
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Network Architecture ◽  
Experimental Tests ◽  
Convergence Time ◽  
Cost Functions ◽  
Great Success ◽  
Feed Forward Neural Network ◽  
Training Time

Artificial Neural Networks are a popular machine learning and artificial intelligence technique, proposed since the 1950s. Among their greatest challenges is the training of parameters such as weights, parameters of the activation functions and constants, as well as their yperparameters, such as network architecture and density of neurons per layer. Among the best known algorithms for parametric optimization of networks are Adam and BP, applied mainly in popular architectures such as MLP, RNN, LSTM, Feed-forward Neural Network (FNN), RBFNN, among many others. Recently, the great success of deep neural networks, known as Deep Learnings, as well as fully connected networks, has faced problems with training time and the use of specialized hardware. These challenges gave new impetus to the use of optimization algorithms for the training of these networks, and more recently to the algorithms inspired by nature, also called as NI. This strategy, although not a recent technique, has not yet received much attention from researchers, requiring today a greater number of experimental tests and evaluation, mainly due to the recent appearance of a much larger range of algorithms NI. Some of the elements that need attention, especially for the most recent NI, are mainly related to the time of convergence and studies on the use of different cost functions. Thus, the present master’s dissertation aims to perform tests, comparisons, and studies on algorithms NI applied to the training of neural networks. Both traditional and recent NI algorithms were tested, from many perspectives, including convergence time and cost functions, elements that until now have received little attention from researchers in previous tests. The results showed that the use of NI algorithms for the training of traditional RNAs obtained results with good classification, similar to popular algorithms such as Adam and BPMA, but surpassing these algorithms in terms of convergence time in 20 up to 70%, depending on the network and the parameters involved. This indicates that the strategy of using NI algorithms, especially the most recent ones, for training neural networks is a promising method that can impact the time and quality of the results of recent and future machine learning applications and artificial intelligence

scholarly journals A primer on deep learning and convolutional neural networks for clinicians

2021 ◽  
Vol 12 (1) ◽  
Author(s):  
Lara Lloret Iglesias ◽  
Pablo Sanz Bellón ◽  
Amaia Pérez del Barrio ◽  
Pablo Menéndez Fernández-Miranda ◽  
David Rodríguez González ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Deep Learning ◽  
Convolutional Neural Networks ◽  
Learning Capabilities ◽  
Computer Vision Applications ◽  
Fully Connected ◽  
Imaging Perspective ◽  
Adaptive Momentum

AbstractDeep learning is nowadays at the forefront of artificial intelligence. More precisely, the use of convolutional neural networks has drastically improved the learning capabilities of computer vision applications, being able to directly consider raw data without any prior feature extraction. Advanced methods in the machine learning field, such as adaptive momentum algorithms or dropout regularization, have dramatically improved the convolutional neural networks predicting ability, outperforming that of conventional fully connected neural networks. This work summarizes, in an intended didactic way, the main aspects of these cutting-edge techniques from a medical imaging perspective.

scholarly journals Gesture Recognition in Images Using Neural Networks

2019 ◽  
Vol 8 (11) ◽  
pp. e278111470
Author(s):  
André Ricardo Nascimento das Neves ◽  
Hugo Kenji Rodrigues Okada ◽  
Ricardo Shitsuka
Keyword(s):  
Artificial Intelligence ◽  
Neural Networks ◽  
Gesture Recognition ◽  
Experimental Tests ◽  
Product Reviews ◽  
Human Reasoning ◽  
Initial Part ◽  
Computer Research

Artificial Intelligence is an area of computer research that is focused on developing mechanisms and devices to simulate human reasoning. Within this, an important subarea is the recognition of images. This article aims to describe the initial part of a research that aims to analyze and identify registered feelings of body expressions in videos of product reviews. Experimental tests have been planned to identify the best technique to solve the problem. Some forms of gesture identification through the use of neural networks were analyzed and tested.

Facial Expression Recognition with Gender Identification

2020 ◽  
pp. 57-63
Author(s):  
admin admin ◽  
◽  
◽  
◽  
◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Facial Expression ◽  
Facial Expression Recognition ◽  
Expression Recognition ◽  
Human Face ◽  
Gender Identification ◽  
Artificial Intelligence Techniques ◽  
The Mind

The human facial emotions recognition has attracted interest in the field of Artificial Intelligence. The emotions on a human face depicts what’s going on inside the mind. Facial expression recognition is the part of Facial recognition which is gaining more importance and need for it increases tremendously. Though there are methods to identify expressions using machine learning and Artificial Intelligence techniques, this work attempts to use convolution neural networks to recognize expressions and classify the expressions into 6 emotions categories. Various datasets are investigated and explored for training expression recognition models are explained in this paper and the models which are used in this paper are VGG 19 and RESSNET 18. We included facial emotional recognition with gender identification also. In this project we have used fer2013 and ck+ dataset and ultimately achieved 73% and 94% around accuracies respectively.

On techniques of expert systems on the example of the Akinator program

2020 ◽  
Vol 25 (2) ◽  
pp. 7-13
Author(s):  
Zhangozha A.R. ◽  
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Expert Systems ◽  
Knowledge Base ◽  
Hybrid Systems ◽  
Online Game ◽  
Basic Principles ◽  
The Moment ◽  
Definition Of

On the example of the online game Akinator, the basic principles on which programs of this type are built are considered. Effective technics have been proposed by which artificial intelligence systems can build logical inferences that allow to identify an unknown subject from its description (predicate). To confirm the considered hypotheses, the terminological analysis of definition of the program "Akinator" offered by the author is carried out. Starting from the assumptions given by the author's definition, the article complements their definitions presented by other researchers and analyzes their constituent theses. Finally, some proposals are made for the next steps in improving the program. The Akinator program, at one time, became one of the most famous online games using artificial intelligence. And although this was not directly stated, it was clear to the experts in the field of artificial intelligence that the program uses the techniques of expert systems and is built on inference rules. At the moment, expert systems have lost their positions in comparison with the direction of neural networks in the field of artificial intelligence, however, in the case considered in the article, we are talking about techniques using both directions – hybrid systems. Games for filling semantics interact with the user, expanding their semantic base (knowledge base) and use certain strategies to achieve the best result. The playful form of such semantics filling programs is beneficial for researchers by involving a large number of players. The article examines the techniques used by the Akinator program, and also suggests possible modifications to it in the future. This study, first of all, focuses on how the knowledge base of the Akinator program is built, it consists of incomplete sets, which can be filled and adjusted as a result of further iterations of the program launches. It is important to note our assumption that the order of questions used by the program during the game plays a key role, because it determines its strategy. It was identified that the program is guided by the principles of nonmonotonic logic – the assumptions constructed by the program are not final and can be rejected by it during the game. The three main approaches to acquisite semantics proposed by Jakub Šimko and Mária Bieliková are considered, namely, expert work, crowdsourcing and machine learning. Paying attention to machine learning, the Akinator program using machine learning to build an effective strategy in the game presents a class of hybrid systems that combine the principles of two main areas in artificial intelligence programs – expert systems and neural networks.

Train Deep Learning Models using subsurface geological images datasets

2021 ◽  
Author(s):  
Ramy Abdallah ◽  
Clare E. Bond ◽  
Robert W.H. Butler
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Deep Learning ◽  
Convolutional Neural Networks ◽  
Learning Models ◽  
Structural Interpretation ◽  
Performance Accuracy ◽  
Learning Techniques ◽  
Machine Learning Models

<p>Machine learning is being presented as a new solution for a wide range of geoscience problems. Primarily machine learning has been used for 3D seismic data processing, seismic facies analysis and well log data correlation. The rapid development in technology with open-source artificial intelligence libraries and the accessibility of affordable computer graphics processing units (GPU) makes the application of machine learning in geosciences increasingly tractable. However, the application of artificial intelligence in structural interpretation workflows of subsurface datasets is still ambiguous. This study aims to use machine learning techniques to classify images of folds and fold-thrust structures. Here we show that convolutional neural networks (CNNs) as supervised deep learning techniques provide excellent algorithms to discriminate between geological image datasets. Four different datasets of images have been used to train and test the machine learning models. These four datasets are a seismic character dataset with five classes (faults, folds, salt, flat layers and basement), folds types with three classes (buckle, chevron and conjugate), fault types with three classes (normal, reverse and thrust) and fold-thrust geometries with three classes (fault bend fold, fault propagation fold and detachment fold). These image datasets are used to investigate three machine learning models. One Feedforward linear neural network model and two convolutional neural networks models (Convolution 2d layer transforms sequential model and Residual block model (ResNet with 9, 34, and 50 layers)). Validation and testing datasets forms a critical part of testing the model’s performance accuracy. The ResNet model records the highest performance accuracy score, of the machine learning models tested. Our CNN image classification model analysis provides a framework for applying machine learning to increase structural interpretation efficiency, and shows that CNN classification models can be applied effectively to geoscience problems. The study provides a starting point to apply unsupervised machine learning approaches to sub-surface structural interpretation workflows.</p>

Stock Market Prediction through Artificial Intelligence, Machine Learning and Neural Networks

2021 ◽  
Author(s):  
Ambarish Shashank Gadgil ◽  
Aditya Fakirmohan Desity ◽  
Prasanna Hemant Asole ◽  
Harsh Shailesh Dandge ◽  
Spurti Shinde
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Stock Market ◽  
Stock Market Prediction

Removing the Constraints of Disability

2021 ◽  
pp. 205-219
Author(s):  
Wadad Kathy Tannous ◽  
Laney McGrew
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Assistive Technologies ◽  
Vision Impairment ◽  
Cultural Participation ◽  
Smart Glasses ◽  
Control Computer ◽  
3D Printed ◽  
Over Time

One billion people globally live with disabilities that are physical, sensory, psychiatric, neurological, cognitive, or intellectual. Their disabilities are dynamic and can be temporary or permanent, singular or plural, from birth or developed, and can change over time. People with disabilities face barriers to economic, social, political, and cultural participation. Assistive technology, artificial intelligence, and broader technology can amplify their inclusion, participation, and independence. This chapter will highlight emerging and evolving technologies, rooted in machine learning and neural networks, which assist across different disabilities and seek to improve the user's sense of ability and independence. These include Seeing AI app, OXSIGHT, OrCam, Envision smart glasses, and Dot Watch for vision impairment; Ava app and cognitive hearing aid for hearing impairment; Liftware self-stabilising utensils for limited hand mobility; Eyegaze and Tobii – assistive technologies that allow users to control computer and smartphone screens with their eyes; and 3D printed prosthetics.

scholarly journals The Interplay of Knowledge Representation with Various Fields of Artificial Intelligence in Medicine

2019 ◽  
Vol 28 (01) ◽  
pp. 027-034 ◽  
Author(s):  
Laszlo Balkanyi ◽  
Ronald Cornet
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Fuzzy Logic ◽  
Knowledge Representation ◽  
Expert Systems ◽  
Frequency Analysis ◽  
Medical Subject Headings ◽  
Subject Headings

Introduction: Artificial intelligence (AI) is widespread in many areas, including medicine. However, it is unclear what exactly AI encompasses. This paper aims to provide an improved understanding of medical AI and its constituent fields, and their interplay with knowledge representation (KR). Methods: We followed a Wittgensteinian approach (“meaning by usage”) applied to content metadata labels, using the Medical Subject Headings (MeSH) thesaurus to classify the field. To understand and characterize medical AI and the role of KR, we analyzed: (1) the proportion of papers in MEDLINE related to KR and various AI fields; (2) the interplay among KR and AI fields and overlaps among the AI fields; (3) interconnectedness of fields; and (4) phrase frequency and collocation based on a corpus of abstracts. Results: Data from over eighty thousand papers showed a steep, six-fold surge in the last 30 years. This growth happened in an escalating and cascading way. A corpus of 246,308 total words containing 21,842 unique words showed several hundred occurrences of notions such as robotics, fuzzy logic, neural networks, machine learning and expert systems in the phrase frequency analysis. Collocation analysis shows that fuzzy logic seems to be the most often collocated notion. Neural networks and machine learning are also used in the conceptual neighborhood of KR. Robotics is more isolated. Conclusions: Authors note an escalation of published AI studies in medicine. Knowledge representation is one of the smaller areas, but also the most interconnected, and provides a common cognitive layer for other areas.

Rule Generation Methods Based on Logic Synthesis

2011 ◽  
pp. 978-983
Author(s):  
Marco Muselli
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Neural Networks ◽  
Real World ◽  
Inductive Reasoning ◽  
Logic Synthesis ◽  
Rule Generation ◽  
Research Areas ◽  
Real World Problem

One of the most relevant problems in artificial intelligence is allowing a synthetic device to perform inductive reasoning, i.e. to infer a set of rules consistent with a collection of data pertaining to a given real world problem. A variety of approaches, arising in different research areas such as statistics, machine learning, neural networks, etc., have been proposed during the last 50 years to deal with the problem of realizing inductive reasoning.

scholarly journals Interaction between Model Based Signal and Image Processing, Machine Learning and Artificial Intelligence

Proceedings ◽  
2019 ◽  
Vol 33 (1) ◽  
pp. 16
Author(s):  
Ali Mohammad-Djafari
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Image Processing ◽  
Bayesian Inference ◽  
Missing Values ◽  
General Purpose ◽  
Parametric Models ◽  
Great Success ◽  
Iterative Regularization ◽  
Model Based

Signale and image processing has always been the main tools in many area and in particular in Medical and Biomedical applications. Nowadays, there are great number of toolboxes, general purpose and very specialized, in which classical techniques are implemented and can be used: all the transformation based methods (Fourier, Wavelets, ...) as well as model based and iterative regularization methods. Statistical methods have also shown their success in some area when parametric models are available. Bayesian inference based methods had great success, in particular, when the data are noisy, uncertain, incomplete (missing values) or with outliers and where there is a need to quantify uncertainties. In some applications, nowadays, we have more and more data. To use these “Big Data” to extract more knowledge, the Machine Learning and Artificial Intelligence tools have shown success and became mandatory. However, even if in many domains of Machine Learning such as classification and clustering these methods have shown success, their use in real scientific problems are limited. The main reasons are twofold: First, the users of these tools cannot explain the reasons when the are successful and when they are not. The second is that, in general, these tools can not quantify the remaining uncertainties. Model based and Bayesian inference approach have been very successful in linear inverse problems. However, adjusting the hyper parameters is complex and the cost of the computation is high. The Convolutional Neural Networks (CNN) and Deep Learning (DL) tools can be useful for pushing farther these limits. At the other side, the Model based methods can be helpful for the selection of the structure of CNN and DL which are crucial in ML success. In this work, I first provide an overview and then a survey of the aforementioned methods and explore the possible interactions between them.

Sign in / Sign up

Export Citation Format

Share Document