The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
This paper presents the design of a convolutional neural network architecture using the MatConvNet library for MATLAB in order to achieve the recognition of 2 classes of hand gestures: ”open” and ”closed”. Six architectures were implemented to which their hyperparameters and depth were varied to observe their behavior through the validation error in the training and accuracy in the estimation of each...
Our work in this paper presents a prediction of quality of experience based on full reference parametric (SSIM, VQM) and application metrics (resolution, bit rate, frame rate) in SDN networks. First, we used DCR (Degradation Category Rating) as subjective method to build the training model and validation, this method is based on not only the quality of received video but also the original video but...
It is well-known that the precision of data, weight vector, and internal representations employed in learning systems directly impacts their energy, throughput, and latency. The precision requirements for the training algorithm are also important for systems that learn on-the-fly. In this paper, we present analytical lower bounds on the precision requirements for the commonly employed stochastic gradient...
We propose a new pretreatment for pedestrian detection with convolutional networks. It is widely known that the phenomenon of overlapping feature distribution is common, which leads to overfitting problem. We present a method that divide one category that have overlapping distributed features into multi-subcategories. By this means smooth boundaries can be easily found to separate different subcategories,...
The data are generated very rapidly from different information sources. These generation of data is increasing day by day from various sources such as automated data collection tools, database systems, e-commerce and social media websites. There is an explosive growth of data from terabytes to petabytes. It is essential to extract valuable knowledge from these large data. Since large amount of data...
Machine Learning plays very important role in processing of large amounts of structured and unstructured data. A set of algorithms can be used to get meaningful insights into the data that are helpful in making effective business decisions. Document clustering is one of the popular machine learning technique used to group unstructured data (text documents) based on its content and further analyze...
There exists a base classification system for classification of problem tickets in the Enterprise domain. Different deep learning algorithms (Gated Recursive Unit and Long Short Term Memory) were investigated for solving the classification problem. Experiments were conducted for different parameters and layers for these algorithms. Paper brings out the architectures tried, results obtained, our conclusions...
The urban classification of PolSAR images is made difficult by the characteristic of a rotated target to exhibit volume scattering. In this paper we use a deep learning technique in conjunction with some statistical parameters to learn to classify urban areas irrespective of the rotation. The learning algorithm was trained to differentiate urban from non-urban areas and was able to achieve a 8.5834%...
Different types of classifiers were investigated in the context of classification of problem tickets in the Enterprise domain. There were still challenges in building an accurate classifier post data cleaning and other accuracy improving pre-processing techniques. Creating an ensemble of classifiers gave better accuracy than individual classifiers. The maximum accuracy was got by enhancing the ensemble...
A new method of classification of a speaker’s gender based on cumulant coefficients is proposed. The effect of an additive noise and measurement error of classification signs on accuracy of classification is analyzed. The expediency of construction of an adaptive system of classification operating with considering of masking of a speech signal by noise is shown. Comparison of the proposed method of...
It is estimated that 80% of crashes and 65% of near collisions involved drivers inattentive to traffic for three seconds before the event. This paper develops an algorithm for extracting characteristics allowing the cell phones identification used during driving a vehicle. Experiments were performed on sets of images with 100 positive images (with phone) and the other 100 negative images (no phone),...
Functional diagnosis for complex electronic boards is a time-consuming task that requires big expertise to the diagnosis engineers. In this paper we propose a new engine for board-level adaptive incremental functional diagnosis based on decision trees. The engine incrementally selects the tests that have to be executed and based on the test outcomes it automatically stops the diagnosis as soon as...
machine learning algorithms are widely used in classification problems. Certainly, recognition quality of algorithms is important indicator, but the ability of the algorithm to learn is more significant. In this work the learning curves experiment was performed in order to identify which of the three learning rates occur when training the machine learning algorithms: overfitting, perfect case and...
Research that explores the use of machine learning for automatic security classification of information objects is about to emerge. In this paper we investigate the opportunity to increase the machine learning performance by taking advantage from time information that is "hidden" in the documents of the training set. This paper presents a technique to do so, and confirms that this is a promising...
Methods for monitoring the human physical activity are recently investigated in order to assess the health status of the individuals and thus promote a healthier lifestyle. This paper proposes ‘dist-colorimetrics’, a methodology that aims to represent and classify ambulatory activities based on the spectral distances measures. A data collection platform including four accelerometer sensors mounted...
Children with neurological disorder such as autism need personalized development system for their daily activities. Technology can play a significant role. This paper introduces a research to deliver personalized learning materials for children with special needs based on diverse characteristics of children. There are four parts of the system: i) identifying level of the user by using machine learning...
This paper tackles the Romanian syllabification and stress assignment problems, and proposes an efficient machine learning based solution. We show that by designing the appropriate feature sets for each specific problem, learning algorithms achieve satisfactory accuracy rates for both problems (∼92% for syllabification, ∼85% for stress assignment), even for relatively small training set sizes. We...
With the increasing risk of data leakage, information guards have emerged as a novel concept in the field of security which bears similarity to spam filter that examine the content of the exchanged messages. A guard is defined as a high-assurance device used to control the information flow, typically from a domain with a "high" level of confidentiality, such as a corporate or military network,...
A research approach of crack detection of rotating shafts based on acoustic emission (AE) signals and machine learning is proposed in this paper. The relationship between crack intensity and domain features are investigated, and the features which could well indicate the crack condition are selected for modelling and crack prediction. Multiple Linear Regression (MLR), Artificial Neural Networks (ANN)...
A method for sentiment polarity assignment for textual content written in Polish using supervised machine learning approach with transfer learning scheme is proposed in the paper. It has been shown that performing simple natural language processing steps prior to classification, provides inspiring results without redundant computation overhead. The documents containing subjective opinions were classified...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.