Generative adversarial networks (GANs) have become popular and powerful models for solving a wide range of image processing problems. We introduce a novel component based on image quality measures in the objective fun...
详细信息
ISBN:
(数字)9783030623623
ISBN:
(纸本)9783030623616;9783030623623
Generative adversarial networks (GANs) have become popular and powerful models for solving a wide range of image processing problems. We introduce a novel component based on image quality measures in the objective function of GANs for solving image deblurring problems. Such additional constraints can regularise the training and improve the performance. Experimental results demonstrate marked improvements on generated or restored image quality both quantitatively and visually. Boosted model performances are observed and testified on three test sets with four image quality measures. It shows that image quality measures are additional flexible, effective and efficient loss components to be adopted in the objective function of GANs.
This paper covers the analysis of contemporary portfolio management techniques, as well as the development of a backtesting trading environment and an autonomous trading agent based on state-of-the-art Reinforcement L...
详细信息
ISBN:
(纸本)9781665424271
This paper covers the analysis of contemporary portfolio management techniques, as well as the development of a backtesting trading environment and an autonomous trading agent based on state-of-the-art Reinforcement learning algorithms with different reward functions. The scientific novelty of this study relies on the development of a scalable automated trading agent that makes periodic investment decisions autonomously. Baseline strategies include UCPR (weekly rebalancing) and UBAH (buy and hold strategy) for SPDR SP 500 ETF Trust (NYSE: SPY). Backtesting Period covers the range 2017-0101-2020-03-30, characterized by the sustained growth of the US stock market in 2017.
At telecommunications companies, call-centers have the highest interaction with customers, and the operators' performance is vital because an excellent service satisfies the customer and helps a better operation. ...
详细信息
ISBN:
(纸本)9783030623616;9783030623623
At telecommunications companies, call-centers have the highest interaction with customers, and the operators' performance is vital because an excellent service satisfies the customer and helps a better operation. Therefore, attempts are made to use customer data, call operator data, and historical service data to improve support. Pairing a customer with an operator who is comfortable with the problem to solve helps companies reducing costs, improves customer service, and increases employee productivity. In this article, we propose an approach based on machine learning and optimization, which predicts the problem for which the customer is calling and routes the call and the customer to the most appropriate call operator. The results show that using large amounts of business data along with innovative algorithms such as LightGBM can improve the customer support performance.
In this paper a software assistant is presented, which supports the users of a small community app in the creation of ads in the social marketplace by pointing out missing information and useful additions. For this pu...
详细信息
ISBN:
(纸本)9789897583728
In this paper a software assistant is presented, which supports the users of a small community app in the creation of ads in the social marketplace by pointing out missing information and useful additions. For this purpose, questions about potentially missing aspects of the content should create incentives to supplement the missing information. An insight into the prototypical development of the software assistant shows that automated support functions can be provided for the users with machine learning procedures and natural language processing even despite data protection restrictions and less data. The focus of this paper is on the presentation of text creation support. Its implementation reveals problems with the use of German language models and their language processing and counteracts these with a rule-based approach. The learning ability of the system through automatedlearning procedures enables the software assistants to react and categorize to linguistic and content-related changes in the input text of the users.
This paper compares three supervised learning techniques (support vector machines, random forest, and decision trees with gradient boosting) applied to the problem of detecting lane change maneuvers of a remote vehicl...
详细信息
ISBN:
(纸本)9781728103235
This paper compares three supervised learning techniques (support vector machines, random forest, and decision trees with gradient boosting) applied to the problem of detecting lane change maneuvers of a remote vehicle using Vehicle-to-Vehicle (V2V) safety messages based on automotive standards. The feature vector used for training is derived from the transmitted measurements smoothed over a sliding window, and it includes some differential measurements. Classifier training and evaluation via cross validation is performed on a real vehicle data set consisting of 740 km of drive data capturing over 1000 lane change maneuvers on California highways. The results show that the supervised learning techniques successfully predict 98.4% of remote vehicle lane changes with a mean detection time of 0.31 sec on straight roadways and 89.5% of lane changes with a mean detection time of 0.62 sec on curved roadways. The detection algorithm has also shown to be robust against packet loss for a wide range of Packet Error Rate (PER).
Lately the government has produced many documentations which analyzed higher education legislation and which maps out the future of higher education. These documents strongly follow the teaching strategies focusing in...
详细信息
ISBN:
(纸本)9781479976775
Lately the government has produced many documentations which analyzed higher education legislation and which maps out the future of higher education. These documents strongly follow the teaching strategies focusing in particular on the e-learningstrategies and the increased need for online training. These documents include specific agenda relating to the training and preparation of teachers, the presentation of curriculum and the development of schemes of teaching and learning. It has political requirement that students be given the opportunity learning in online environment. As adults, they will only be able to keep up with the challenge of global knowledge exchange and be able to use interactive elements.
The Personal data Protection Act (PDPA) was created to prevent the breach of personal information of users of computer systems without the data owner's consent. One type of data that frequently has problems with p...
详细信息
ISBN:
(纸本)9798350381771;9798350381764
The Personal data Protection Act (PDPA) was created to prevent the breach of personal information of users of computer systems without the data owner's consent. One type of data that frequently has problems with privacy violations is images and videos. Because of difficult control, as a result, there are often extraneous people in the frame instead of just the intended subject. If the person caught in the frame does not want this information published, there will be a problem with that video. This causes the identity of the person to be concealed so that they can be identified. Although doing this is an acceptable method, censorship is labor-intensive and time-consuming. For these reasons, we proposed the automated Face Selection and Censoring on Image and Video System using Multi-Task Cascaded Convolutional Neural Networks (MTCNN) Model to automatically detect the face and censor only unwanted persons. Furthermore, our proposed method also executes an automated system that can sense and ignore some frames that are not essential or redundant with other nearby frames to reduce complex processing and time-consumption.
The aim of this paper is to identify and understand bot activity in twitter discussion. The prevalence of Twitter bots have gained significant limelight recently due to their misuse in influencing public sentiment for...
详细信息
ISBN:
(纸本)9781450387910
The aim of this paper is to identify and understand bot activity in twitter discussion. The prevalence of Twitter bots have gained significant limelight recently due to their misuse in influencing public sentiment for political gains. For our analysis, we use Twitter data of 2019 Canadian Elections. We perform principal component analysis and K-means clustering on the data set. Using the results we isolate bots from human accounts.
The average annual daily bicyclist volume (AADB) is a measure being used in non-motorized transportation studies such as exposure modeling. This metric can be estimated by averaging the daily bicyclist's volume me...
详细信息
ISBN:
(纸本)9781728103235
The average annual daily bicyclist volume (AADB) is a measure being used in non-motorized transportation studies such as exposure modeling. This metric can be estimated by averaging the daily bicyclist's volume measured throughout the year with a long-term automated counter. As continuous data for a whole year at many sites may not be available, a common practice is to collect short-term count data for a sample of locations and then apply an extrapolation method to convert short-term count data to yearly count data. To perform extrapolation, each short-term counter must be matched to one or more long-term counters with similar demand patterns. The matching procedure can significantly impact the accuracy of AADB estimation. This study proposes a matching approach based on a cluster analysis approach, Partitioning Around Medoids (PAM), and a supervised learning approach, K-Nearest Neighbor (KNN). It was found that the proposed approach using a combination of certain variables such as land use and a traffic distribution index resulted in low AADB estimation error rates, which enhanced existing approaches.
The combination of machine learning techniques and signal analysis is a well-known solution for the fault diagnosis of industrial equipment. Efficient maintenance management, safer operation, and economic gains are th...
详细信息
ISBN:
(数字)9783030623623
ISBN:
(纸本)9783030623616;9783030623623
The combination of machine learning techniques and signal analysis is a well-known solution for the fault diagnosis of industrial equipment. Efficient maintenance management, safer operation, and economic gains are three examples of benefits achieved by using this combination to monitor the equipment condition. In this context, the selection of meaningful information to train machine learning models arises as an important issue, since it influences the model accuracy and complexity. Aware of this, we propose to use the ratio between the interclass and intraclass Kullback-Leibler divergence to identify promising data for training fault diagnosis models. We assessed the performance of this metric on compressor fault datasets. The results suggested a relation between the model accuracy and the ratio between the average interclass and intraclass divergences.
暂无评论