The paper presents research in usability of web technologies for implementation of machine learning and clustering algorithms into embedded systems. The research work is divided into two main parts. The first part is ...
详细信息
ISBN:
(纸本)9781479976775
The paper presents research in usability of web technologies for implementation of machine learning and clustering algorithms into embedded systems. The research work is divided into two main parts. The first part is devoted to designing backend system with fast C++ application for learning execution model. The second part of application is frontend based web application with PHP and AJAX to provide interface for virtual laboratory access via internet. This solution is implemented and tested on selected embedded systems (Orange PI Lite, Raspberry PI3).
Automatic speaker assignment can enhance the efficiency of automated systems used for creating meeting minutes. Therefore, various methods of speaker identification have been widely utilized;however, in such methods, ...
详细信息
ISBN:
(纸本)9781728197326
Automatic speaker assignment can enhance the efficiency of automated systems used for creating meeting minutes. Therefore, various methods of speaker identification have been widely utilized;however, in such methods, it is necessary to optimally arrange the equipment. In this paper, we propose a method for identifying speakers using an omnidirectional camera and a microphone. In this process, tens of minutes of image and voice data were used for model training, and the speaker was successfully identified. In the speaker discrimination experiments, the proposed method was able to identify the speaker with an average success rate of more than 70.0% using the lip height and product of the lip height feature and lip width feature as the lip movement features.
datasets for the classification task are usually encoded by a matrix of numbers, the order of rows and columns does not matter. Swapping any two objects or features in it does not change the hidden target function and...
详细信息
ISBN:
(纸本)9783030623647;9783030623654
datasets for the classification task are usually encoded by a matrix of numbers, the order of rows and columns does not matter. Swapping any two objects or features in it does not change the hidden target function and performance of the machine learning algorithms train of the dataset. However, in the dataset generation problem solution such symmetry is an obstacle. In this paper, we study several methods of the inverse transformation of classification dataset aiming to break the symmetry. We experimented with it in the meta-learning problems of datasets generation and algorithm selection which were solved by conditional generative adversarial nets with convolutional networks.
To bring autonomous driving onto public roads, autonomous vehicles must be able to make safe driving decisions. In order to achieve this, they need to be self-aware, meaning they have to be aware of their current capa...
详细信息
ISBN:
(纸本)9781728103235
To bring autonomous driving onto public roads, autonomous vehicles must be able to make safe driving decisions. In order to achieve this, they need to be self-aware, meaning they have to be aware of their current capabilities at all times. One step towards self-awareness is the assessment of the quality of the available sensor data and the estimation of its impact on the processing chain. Knowing when sensor data is compromised will contribute to safer driving decisions of the vehicle. In this contribution, we present a novel, deep-learning approach for overexposure detection in camera images as one step towards sensor data quality monitoring.
Driving style (DS) classification and identification plays an increasingly important role in the development of advanced driver assistance systems and automated vehicles. Both the enhancement of driving safety and the...
详细信息
ISBN:
(纸本)9781728103235
Driving style (DS) classification and identification plays an increasingly important role in the development of advanced driver assistance systems and automated vehicles. Both the enhancement of driving safety and the improvement of fuel efficiency are essential goals of current research in driving style characterization. However, the comfort perspective has still hardly been investigated, despite its importance for the future of driving automation. This paper proposes a driving style classification method, focused on global comfort of the driver and the passengers, but which can also be integrated into the above safety-efficiency viewpoint. Although human comfort in vehicles is affected by different factors, the amplitude and frequency of accelerations are recognized as key signals for assessing driving comfort. The proposed DS classification approach is based on a hybrid machine learning method that combines an unsupervised clustering method with a data-driven extreme learning machine (ELM) algorithm. Hierarchical clustering is used to explore the relevance of the acceleration components in relation to ride comfort, while a single layer ELM topology is implemented to model the DS classifier. The method has been evaluated using experimental data obtained with an instrumented car equipped with in-vehicle sensors and measurement units. The obtained clustering results are consistent with comfort standard indicators, while the data driven algorithm provides encouraging results: more than 95% classification rate using unseen data.
The growing number of intelligent components inside a car leads to a considerable increase in amount of the produced data. Context aware paradigm plays a major role in managing this data and offering a numerous number...
详细信息
ISBN:
(纸本)9781728103235
The growing number of intelligent components inside a car leads to a considerable increase in amount of the produced data. Context aware paradigm plays a major role in managing this data and offering a numerous number of prospects and advantages for existing and new intelligent applications inside the car. Following that, enabling context prediction promises reliable solutions in terms of enhancing the comfort of the occupants and vehicle dynamics. Moreover, this would be a great step toward facilitating highly automated and autonomous driving. However, due to the complex nature of the data resources in an intelligent car and also the lack of comprehensive studies on different aspects of this concept in automotive, defining a functional architecture for context prediction requires broad knowledge and better understanding of multiple domains which are involved and have impacts. In this paper, we investigate the most effective elements and factors in each one of the related domains which help to enable context prediction architectures inside the intelligent cars and analyze the feasible dimensions in detail, cover their advantages, and address the challenges ahead. We elucidate the possibility and validity of our considerations with the help of two use cases of adaptive HVAC and ACC systems.
The use of sensor technology constantly gathering aircrafts' status data has promoted the rapid development of data-driven solutions in aerospace engineering. These methods assist, for instance, with determining a...
详细信息
ISBN:
(纸本)9781728103235
The use of sensor technology constantly gathering aircrafts' status data has promoted the rapid development of data-driven solutions in aerospace engineering. These methods assist, for instance, with determining appropriate actions for aircraft maintenance, repair and overhaul (MRO). Challenges however are found when dealing with such large amounts of data. Identifying patterns, anomalies and faults disambiguation, with acceptable levels of accuracy and reliability are examples of complex problems in this area. Experiments using deep learning techniques, however, have demonstrated its usefulness in assisting on the analysis aircraft health data. The purpose of this paper therefore is to conduct a survey on deep learning architectures and their application in aircraft MRO. Although deep learning in general is not yet largely exploited for aircraft health, from our search, we identified four main architectures employed to MRO, namely, Deep Autoencoders, Long Short-Term Memory, Convolutional Neural Networks and Deep Belief Networks. For each architecture, we review their main concepts, the types of problems to which these architectures are employed to, the type of data used and their outcomes. We also discuss how research in this area can be advanced by identifying current research gaps and outlining future research opportunities.
Taxpayers may be interested in overpayment and which group of taxpayers he or she belongs to. Government officials may be concerned with underpaying taxpayers for auditing purposes and group taxpayers in the rapidly c...
详细信息
ISBN:
(纸本)9781450387910
Taxpayers may be interested in overpayment and which group of taxpayers he or she belongs to. Government officials may be concerned with underpaying taxpayers for auditing purposes and group taxpayers in the rapidly changing society. Machine learning and data mining techniques have been applied to provide solutions to these taxation related queries. Classification algorithms allow predicting the tax bracket based on the taxpayers' attributes. The regression model allows to predict the tax estimate so that the overpayment or underpayment can be determined. Clustering algorithms group taxpayers so that they can be compared to the past year tax brackets. Finally, feature selection allows finding salient attributes to predict the tax and tax bracket. In this article, New York state's Open Tax data is used to demonstrate the machine learning and data mining algorithms and identify issues of using them. Furthermore, various visualization techniques are to present the discovered information to both taxpayers and government officials.
We investigate how automated, data-driven, personalized feedback in a large-scale intelligent tutoring system (ITS) improves student learning outcomes. We propose a machine learning approach to generate personalized f...
详细信息
ISBN:
(纸本)9783030522407;9783030522391
We investigate how automated, data-driven, personalized feedback in a large-scale intelligent tutoring system (ITS) improves student learning outcomes. We propose a machine learning approach to generate personalized feedback, which takes individual needs of students into account. We utilize state-of-the-art machine learning and natural language processing techniques to provide the students with personalized hints, Wikipedia-based explanations, and mathematical hints. Our model is used in Korbit (https://***), a large-scale dialogue-based ITS with thousands of students launched in 2019, and we demonstrate that the personalized feedback leads to considerable improvement in student learning outcomes and in the subjective evaluation of the feedback.
In this paper we propose a deep learning approach to analyzing driving styles from multivariate timeseries of driving behaviors, a deep topological map of vehicle maneuvers. Our neural network learns a topological map...
详细信息
ISBN:
(纸本)9781728103235
In this paper we propose a deep learning approach to analyzing driving styles from multivariate timeseries of driving behaviors, a deep topological map of vehicle maneuvers. Our neural network learns a topological map that embeds and clusters short subsequences of driving behavior data in a latent space. Our approach is tested on a highway-driving dataset that contains driving behavior data recorded with 59 drivers using a driving simulator. The experiments demonstrated that our proposed method embedded the driving behavior data into a two-dimensional, topological map where extracted elemental driving behaviors are represented as clusters. Individual driving episodes can be compactly represented as estimated probabilistic distributions of elemental driving behaviors over the learned map. We further demonstrated that clusters of driving episodes, "driving styles", and their related features can be efficiently inferred in a data-driven manner by the application of a nonparametric Bayesian method to these estimated probabilistic distributions.
暂无评论