956 results on '"Training phase"'
Search Results
2. Pseudo-model-free hedging for variable annuities via deep reinforcement learning.
- Author
-
Chong, Wing Fung, Cui, Haoen, and Li, Yuxuan
- Subjects
DEEP reinforcement learning ,REINFORCEMENT learning ,VARIABLE annuities ,REWARD (Psychology) ,HEDGING (Finance) ,SEQUENTIAL learning - Abstract
This paper proposes a two-phase deep reinforcement learning approach, for hedging variable annuity contracts with both GMMB and GMDB riders, which can address model miscalibration in Black-Scholes financial and constant force of mortality actuarial market environments. In the training phase, an infant reinforcement learning agent interacts with a pre-designed training environment, collects sequential anchor-hedging reward signals, and gradually learns how to hedge the contracts. As expected, after a sufficient number of training steps, the trained reinforcement learning agent hedges, in the training environment, equally well as the correct Delta while outperforms misspecified Deltas. In the online learning phase, the trained reinforcement learning agent interacts with the market environment in real time, collects single terminal reward signals, and self-revises its hedging strategy. The hedging performance of the further trained reinforcement learning agent is demonstrated via an illustrative example on a rolling basis to reveal the self-revision capability on the hedging strategy by online learning. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
3. Rotational invariant biologically inspired object recognition
- Author
-
Hiwa Sufi karimi and Karim Mohammadi
- Subjects
object recognition purposes ,robust method ,image variations ,location changes ,applied images ,training phase ,Photography ,TR1-1050 ,Computer software ,QA76.75-76.765 - Abstract
The biologically inspired hierarchical model and X (HMAX) has been one of the superior techniques for object recognition purposes. HMAX is a robust method in the presence of some image variations including illumination, different scales, and location changes. However, the performance of HMAX extremely deteriorates if the orientation of the applied images in training phase is different than the orientation of testing images. In this study, the authors propose rotational invariant HMAX (RIMAX) to overcome the existing issues in object recognition imposed by rotations in the images. To this end, they embed two new layers into the structure of the standard HMAX. In the first added layer, non‐accidental properties (e.g. corners and edges) are extracted as features that lead to obtaining a repeatable object recognition process. The second added layer provides robustness of RIMAX to image rotations by normalising the dominant orientation of the extracted features. Furthermore, they considerably reduce the imposed computational load by modifying template matching strategy as well as removing multiple scales of the Gabor filter. Simulation results employ Caltech101, TUD, Caltech5, and GRAZ‐02 databases that validate RIMAX outperforms the standard HMAX and the other mathematical approaches in terms of robustness, accuracy, repeatability, and speed.
- Published
- 2020
- Full Text
- View/download PDF
4. 竞速类项目的训练负荷强度分布一一区别与规律.
- Author
-
陈亮
- Abstract
Copyright of Journal of Tianjin University of Sport / Tianjin Tiyu Xueyuan Xuebao is the property of Tianjin University of Sport and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2022
- Full Text
- View/download PDF
5. Case Study: Interpretability of Fuzzy Systems Applied to Identity Verification
- Author
-
Cpałka, Krzysztof, Kacprzyk, Janusz, Series editor, and Cpałka, Krzysztof
- Published
- 2017
- Full Text
- View/download PDF
6. Comparison of Optimization Techniques for Modular Neural Networks Applied to Human Recognition
- Author
-
Sánchez, Daniela, Melin, Patricia, Carpio, Juan, Puga, Hector, Kacprzyk, Janusz, Series editor, Melin, Patricia, editor, and Castillo, Oscar, editor
- Published
- 2017
- Full Text
- View/download PDF
7. Image enhancement using convolutional neural network to identify similar patterns.
- Author
-
Lu, Ching‐Ta, Chen, Ruei‐Han, Wang, Ling‐Ling, and Lin, Jia‐An
- Abstract
An image may be disturbed by impulse noise during transmission or acquisition. To effectively restore the disturbed image is important for the applications of image processing. This study aims at enhancing the disturbed images by using the convolutional neural network (CNN) to identify similar patterns for the restoration of noisy pixels. In the training phase, each noisy pixel is analysed and compared with the noise‐free image to find the closest neighbouring pixels. The pixels in a local window form a micro‐pattern. All the captured micro‐patterns, whose centre pixel is noisy, become a dataset for the training of a position CNN. The closest neighbouring pixel of a noisy image to the centre one of the noise‐free image at the same position of each micro‐pattern is selected to be the target. In the enhancement phase, a noisy micro‐pattern, where the centre pixel is noisy, is input into the trained position CNN. The top N pixels are recognised and averaged to replace the grey level of the centre pixel. An enhanced pixel is obtained. The experimental results show that the position CNN can well recognise the similar neighbouring pixels and effectively enhance the noisy pixels in an image disturbed by salt‐and‐pepper noise. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
8. Rotational invariant biologically inspired object recognition.
- Author
-
Sufi karimi, Hiwa and Mohammadi, Karim
- Abstract
The biologically inspired hierarchical model and X (HMAX) has been one of the superior techniques for object recognition purposes. HMAX is a robust method in the presence of some image variations including illumination, different scales, and location changes. However, the performance of HMAX extremely deteriorates if the orientation of the applied images in training phase is different than the orientation of testing images. In this study, the authors propose rotational invariant HMAX (RIMAX) to overcome the existing issues in object recognition imposed by rotations in the images. To this end, they embed two new layers into the structure of the standard HMAX. In the first added layer, non‐accidental properties (e.g. corners and edges) are extracted as features that lead to obtaining a repeatable object recognition process. The second added layer provides robustness of RIMAX to image rotations by normalising the dominant orientation of the extracted features. Furthermore, they considerably reduce the imposed computational load by modifying template matching strategy as well as removing multiple scales of the Gabor filter. Simulation results employ Caltech101, TUD, Caltech5, and GRAZ‐02 databases that validate RIMAX outperforms the standard HMAX and the other mathematical approaches in terms of robustness, accuracy, repeatability, and speed. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
9. Neural network‐based image quality comparator without collecting the human score for training.
- Author
-
Bao, Long, Panetta, Karen, and Agaian, Sos
- Abstract
Emulating human behaviours in automated image quality assessment (IQA) enables a comparator framework to remove the differences in human bias naturally. Based on the observation of the practical applications of IQA, this study focuses on similar‐content image quality comparison based on a new image quality comparator (IQC). Outstanding proven IQAs can be utilised in this comparator to achieve a new non‐linear combination strategy to boost the IQAs' performance in image quality comparison. For both input images to be compared, proven IQAs are utilised to obtain nine features from each image, yielding 18 total features. Then, a four‐layer comparison network conducts a classification task to indicate which input image has better quality. In the training phase, the commonly used human scores as training labels are replaced with pairwise comparison results that are automatically generated from assigned distortion level differences. By not utilising human score in training phase, this IQC shows two advantages: (i) it removes huge labor and time cost to collect the human scores and (ii) it solves the problem of over‐fitting benefiting from simplicity of creating a large image training dataset. Furthermore, the experimental tests and cross‐dataset validation comparison tests demonstrate its impressive performance. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
10. Enhanced Few-Shot Learning for Intrusion Detection in Railway Video Surveillance
- Author
-
Xi Chen, Xiao Gong, Zhangdui Zhong, and Wei Chen
- Subjects
Similarity (geometry) ,Computer science ,business.industry ,Mechanical Engineering ,Shot (filmmaking) ,Frame (networking) ,Supervised learning ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,Small sample ,Intrusion detection system ,Track (rail transport) ,Computer Science Applications ,Automotive Engineering ,Training phase ,Computer vision ,Artificial intelligence ,business - Abstract
Video surveillance is gaining increasing popularity to assist in railway intrusion detection in recent years. However, efficient and accurate intrusion detection remains a challenging issue due to: (a) limited sample number: only small sample size (or portion) of intrusive video frames is available; (b) high inter-scene dissimilarity: various railway track area scenes are captured by cameras installed in different landforms; (c) high intra-scene similarity: the video frames captured by an individual camera share a same background. In this paper, an efficient few-shot learning solution is developed to address the above issues. In particular, an enhanced model-agnostic meta-learner is trained using both the original video frames and segmented masks of track area extracted from the video. Moreover, theoretical analysis and engineering solutions are provided to cope with the highly similar video frames in the meta-model training phase. The proposed method is tested on realistic railway video dataset. Numerical results show that the enhanced meta-learner successfully adapts unseen scene with only few newly collected video frame samples, and its intrusion detection accuracy outperforms that of the standard randomly initialised supervised learning.
- Published
- 2022
- Full Text
- View/download PDF
11. Radar high-resolution range profile recognition via multi-SV method
- Author
-
Long Li, Jie Jiang, and Zheng Liu
- Subjects
image classification ,radar target recognition ,support vector machines ,feature extraction ,radar resolution ,radar high-resolution range profile recognition ,ground compound environment ,essential processes ,HRRP target recognition ,classification performance improvement ,novel target recognition method ,multiple support vector method ,training phase ,multiple correlate SV model ,multitarget feature space ,central-coherence distributed regions ,multiple regions ,outlier discrimination ,targets classification ,practical radar target recognition task ,Engineering (General). Civil engineering (General) ,TA1-2040 - Abstract
The radar high-resolution range profile (HRRP) is of great significance for target recognition in the ground compound environment. As two essential processes of HRRP target recognition, discrimination and classification performance improvement are studied in this study. A novel target recognition method is designed, denoted by multiple support vector (SV) method. In the training phase, a multiple correlate SV model is constructed, which separates the multi-target feature space into several central-coherence distributed regions. Each region has been described by the modified SV method, and the category determination is based on the description of multiple regions. This method can overcome the issues of both outlier discrimination and targets classification. Due to the hierarchical process in the testing phase, this method needs less memory and calculation to fit the practical radar target recognition task. Finally, the experiment based on the measured data verifies the excellent performance of this method.
- Published
- 2019
- Full Text
- View/download PDF
12. Radar high-resolution range profile recognition via multi-SV method.
- Author
-
Li, Long, Jiang, Jie, and Liu, Zheng
- Subjects
HIGH resolution imaging ,OUTLIER detection ,RADAR targets ,RADAR target recognition ,SUPPORT vector machines - Abstract
The radar high-resolution range profile (HRRP) is of great significance for target recognition in the ground compound environment. As two essential processes of HRRP target recognition, discrimination and classification performance improvement are studied in this study. A novel target recognition method is designed, denoted by multiple support vector (SV) method. In the training phase, a multiple correlate SV model is constructed, which separates the multi-target feature space into several central-coherence distributed regions. Each region has been described by the modified SV method, and the category determination is based on the description of multiple regions. This method can overcome the issues of both outlier discrimination and targets classification. Due to the hierarchical process in the testing phase, this method needs less memory and calculation to fit the practical radar target recognition task. Finally, the experiment based on the measured data verifies the excellent performance of this method. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
13. A technique for fast physical training of qualified weightlifters at the training stage of preparation.
- Author
-
CHERNOGOROV, DMITRY N., TUSHER, YURI L., BELYAEV, VASILY S., GROSS, EUGENIA R., and KUZNETSOV, SERGEY V.
- Abstract
This article discusses the issue of improving the methodology of special physical training of qualified weightlifters from different weight categories at the training stage of preparation (4-5th year of study). The ratio of the training load during the annual training cycle is presented, and the influence of the developed methodology is determined on the basis of the obtained results on the preparation of weightlifters. Research objective: The objective of thisresearchwas to develop a method of special physical training for qualified weightlifters for the training stage. Research methods and organization: The research took place at the sports school of the Olympic reserve (ORSS) base of the Moscow weightlifting school. The study was conducted during the 2017-2018 school year. A total of 21 athletes participated in this study (14 athletes from the lightweight category and 7 from the heavyweight category), who were in their 4-5th year of training. Research results and conclusions: the obtained research results on the level of sports skill during the weightlifting training preparation phase in the snatch, jerk, and the total sum, which were determined using biased growth, indicatean average rate of 28%. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
14. Dynamics of the snatch technique cinematic parameters in qualified female weightlifters during different periods of training macrocycle.
- Author
-
TOVSTONOH, OLEXANDR, ROZTORHUI, MARIIA, PITYN, MARYAN, PASICHNYK, VICTORIA, MELNYK, VALERYI, ZAHURA, FEDIR, and POPOVYCH, OLEXANDR
- Abstract
The requirements to modern sports necessitate the study of the dynamics of snatch technique parameters during different macrocycle periods of qualified female weightlifters training. The purpose of the study is to investigate the dynamics of changes of the snatch informative kinematic parameters in qualified female weightlifters under training conditions during intermediate, preparatory and competitive stages of training macrocycle. It has been found that such kinematic parameters of female weightlifters movements as the maximum lift height (Hmax), the lift height at maximum speed (H at Vmax) and the clean height in squat (Hclean) are undoubtedly changing in qualified female weightlifters of different groups of weight categories during intermediate, preparatory and competitive stages of training macrocycle. The identified changes in the snatch technique kinematic parameters of female weightlifters during different stages of macrocycle training enable it to monitor the technique proficiency level and to verify the appropriateness of the selected means of qualified female weightlifters training for annual major and qualifying competitions. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
15. Foreground Refinement Network for Rotated Object Detection in Remote Sensing Images
- Author
-
Peng Zhu, Tianyang Zhang, Puhua Chen, Licheng Jiao, Xu Tang, Xiangrong Zhang, and Chen Li
- Subjects
Relation (database) ,Remote sensing (archaeology) ,Feature (computer vision) ,Computer science ,Detector ,General Earth and Planetary Sciences ,Training phase ,Electrical and Electronic Engineering ,Rotation (mathematics) ,Object detection ,Field (computer science) ,Remote sensing - Abstract
Object detection has been a fundamental task in the field of remote sensing and has made considerable progress in recent years. However, the high background complexity in remote sensing images (RSIs) remains challenging. In this article, we propose a refined rotation detector, namely, the Foreground Refinement Network (FoRDet), to alleviate the above problem by leveraging the information of foreground regions from the perspectives of feature and optimization. Specifically, we propose a foreground relation module (FRL) that aggregates the foreground-contextual representations from the coarse stage and improves the discrimination of foreground regions on feature maps in the refined stage. Besides, considering the risk of the potential foreground anchors being overwhelmed in the training phase, we design a foreground anchor reweighting (FRW) loss that integrates the classification confidence and localization accuracy of each foreground anchor from the coarse stage to dynamically regulate their contributions in the refined stage, which highlights the potential foreground anchors. The comprehensive experimental results on three public datasets for rotated object detection DOTA, HRSC2016, and UCAS-AOD demonstrate the effectiveness of our proposed method.
- Published
- 2022
- Full Text
- View/download PDF
16. Developing Deep Learning Models for Storm Nowcasting
- Author
-
Joaquin Cuomo and V. Chandrasekar
- Subjects
Nowcasting ,Standardization ,business.industry ,Computer science ,Deep learning ,media_common.quotation_subject ,Storm ,Machine learning ,computer.software_genre ,law.invention ,law ,General Earth and Planetary Sciences ,Training phase ,Artificial intelligence ,Electrical and Electronic Engineering ,Radar ,business ,Function (engineering) ,Implementation ,computer ,media_common - Abstract
Storm nowcasting relies on reasonably fast sampled radar data, and deep learning (DL) can be used to harness this vast amount of data. Despite all the publications on this topic over the past five years, there are still ad hoc assumptions and a lack of standardization. This work addresses aspects that have not yet been analyzed on the development of DL models for nowcasting systems, such as the effects of different history lengths or using non-convex metrics during the training phase. For example, we show that even if the loss function is varied, it does not significantly influence the predictions, and that the number of predicted frames has a significant impact. We used the experiments' results to propose different models and compare their performance against other DL models. The results show that the proposed models outperform, in many aspects, the existing implementations.
- Published
- 2022
- Full Text
- View/download PDF
17. Survey-Based Location Systems
- Author
-
Gentile, Camillo, Alsindi, Nayef, Raulefs, Ronald, Teolis, Carole, Gentile, Camillo, Alsindi, Nayef, Raulefs, Ronald, and Teolis, Carole
- Published
- 2013
- Full Text
- View/download PDF
18. An ultra-fast bi-phase advanced network for segmenting crop plants from dense weeds
- Author
-
Xiaoguang Zhang, Zhe Chen, Nan Li, and Xiang Liu
- Subjects
Computer science ,Intersection (set theory) ,business.industry ,Feature extraction ,Phase (waves) ,Soil Science ,Pattern recognition ,Market segmentation ,Fusion tree ,Control and Systems Engineering ,Code (cryptography) ,Training phase ,Ultra fast ,Artificial intelligence ,business ,Agronomy and Crop Science ,Food Science - Abstract
Fast and accurate perception of crop plants is critical for robotic weeding. However, due to the similar appearance, it remains challenging to precisely and efficiently segment crop plants from weeds, especially when the weed density is high. To tackle this problem, we devise an ultra-fast bi-phase advanced network (BA-Net) architecture that advances both the feature extraction phase and the model training phase. A bi-path fusion tree structure and a bi-polar distributed loss are introduced for the two phases, respectively. The bi-path fusion tree merges features from two adjacent convolutional stages at each step and efficiently models the scene through a tree-like structure. The bi-polar distributed loss strengthens the training at ambiguous areas, which can greatly enhance the distinctness at detailed structures like the crop plant boundaries. Experiments demonstrate the superiority of the proposed BA-Net to other state-of-the-art methods in terms of both accuracy and efficiency. BA-Net can accurately segment the crop plants from the weedy background, achieving an intersection over union score of 0.963, a F-measure score of 0.981, and a mean absolute error of 0.00376. Also, BA-Net is lightweight with only 0.55 M trainable parameters and achieves ultra-fast speed, i.e., over 340 FPS on a desktop and over 26 FPS on a Jetson TX2 embedded computer. The code of BA-Net is available at https://github.com/ZhangXG001/BA-Net .
- Published
- 2021
- Full Text
- View/download PDF
19. Application of computer‐generated images to train pattern recognition used in semiquantitative immunohistochemistry scoring
- Author
-
Matthias Schmid and Glen Kristiansen
- Subjects
Microbiology (medical) ,Pathology, Clinical ,Diagnostic Tests, Routine ,business.industry ,Computer-generated imagery ,Pattern recognition ,General Medicine ,Immunohistochemistry ,Pattern Recognition, Automated ,Pathology and Forensic Medicine ,Absolute deviation ,Image Interpretation, Computer-Assisted ,Pattern recognition (psychology) ,Image Processing, Computer-Assisted ,Humans ,Immunology and Allergy ,Training phase ,Medicine ,Artificial intelligence ,Computer aided learning ,business ,Software - Abstract
This study aimed to clarify whether the pattern recognition involved in scoring proliferation fractions can be trained by abstract computerized images of virtual tissues. Twenty computer-generated images with randomly distributed blue or red dots were scored by 12 probands (all co-workers or collaborators of the Institute of Pathology, University of Bonn). Afterward, the probands underwent a training phase during which they received an immediate feedback on the actual rate of positivity after each image. Finally, the initial testing series was rescored. In a second round with 15 different probands, 20 Ki-67 immunohistochemistry images of tonsil tissue were scored, followed by the same training phase with computer-generated images, before the immunohistochemistry slides were scored again. Paired t-tests were used to compare the differences in mean rates pre- and post-training. Concerning computerized images, untrained probands scored the percentages of positive dots with a mean deviation from the true rates of 8.2%. Following training, the same testing series was scored significantly better with a mean deviation of 4.9% (mean improvement 3.3%, p
- Published
- 2021
- Full Text
- View/download PDF
20. Improving Online ECG Interpretation Through Self-Generation of Diagnoses During Practice: A Randomized Study
- Author
-
Lorne J. Gula, Sarah Blissett, and Pavel Antiperovitch
- Subjects
Male ,medicine.medical_specialty ,Self generation ,Students, Medical ,education ,Cardiology ,030204 cardiovascular system & hematology ,law.invention ,Education, Distance ,Electrocardiography ,03 medical and health sciences ,0302 clinical medicine ,Randomized controlled trial ,law ,medicine ,Humans ,Prospective Studies ,030212 general & internal medicine ,Medical diagnosis ,Internet ,business.industry ,Large effect size ,Significant difference ,Test (assessment) ,Physical therapy ,Training phase ,Female ,Clinical Competence ,Educational Measurement ,Cardiology and Cardiovascular Medicine ,business - Abstract
Although electrocardiography (ECG) is a fundamental skill for most physicians, trainees have poor diagnostic performance when interpreting ECGs. In this study, we examine a strategy to improve learning ECG interpretation: self-generation of diagnoses during online practice. We randomly assigned medical students and residents to one of 2 ECG interpretation training formats: multiple-choice (MCQ) or self-generation (SG) format, where participants free-text type their diagnosis aided by an autocomplete feature. The training phase consisted of 30 ECGs, after which participants completed an immediate post test and delayed post test (3-4 weeks later). Forty-eight participants completed the training module, 45 completed the immediate post test, and 27 completed the delayed post test. Participants assigned to the SG format scored higher on the immediate post test compared with those who practiced with the MCQ format, with a large effect size (78% vs 57%; d = 0.94; P = 0.02). There was a trend favouring SG on the delayed post test, with a moderate effect size (67% vs 56%; d = 0.65; P = 0.09). However, only 60% of participants completed the delayed post test, which hindered the detection of a statistically significant difference. The SG group made the correct primary diagnosis at a faster rate (32 vs 56 seconds; P0.001) but had a lower detection of secondary diagnoses (22 vs 42%; P = 0.007). Practicing ECG interpretation using self-generation of diagnoses improved immediate post test performance and fluency. Replication in other contexts and with other populations is required to confirm our findings and to further study retention.
- Published
- 2021
- Full Text
- View/download PDF
21. On the Biological Plausibility of Artificial Metaplasticity
- Author
-
Andina, Diego, Ropero-Peláez, Javier, Hutchison, David, Series editor, Kanade, Takeo, Series editor, Kittler, Josef, Series editor, Kleinberg, Jon M., Series editor, Mattern, Friedemann, Series editor, Mitchell, John C., Series editor, Naor, Moni, Series editor, Nierstrasz, Oscar, Series editor, Pandu Rangan, C., Series editor, Steffen, Bernhard, Series editor, Sudan, Madhu, Series editor, Terzopoulos, Demetri, Series editor, Tygar, Doug, Series editor, Vardi, Moshe Y., Series editor, Weikum, Gerhard, Series editor, Ferrández, José Manuel, editor, Álvarez Sánchez, José Ramón, editor, de la Paz, Félix, editor, and Toledo, F. Javier, editor
- Published
- 2011
- Full Text
- View/download PDF
22. Remembering Under the Influence of Unconscious Expectations
- Author
-
Whittlesea, Bruce W. A., Higham, Philip A., editor, and Leboe, Jason P., editor
- Published
- 2011
- Full Text
- View/download PDF
23. Sequential Reduction Algorithm for Nearest Neighbor Rule
- Author
-
Raniszewski, Marcin, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Doug, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Bolc, Leonard, editor, Tadeusiewicz, Ryszard, editor, Chmielewski, Leszek J., editor, and Wojciechowski, Konrad, editor
- Published
- 2010
- Full Text
- View/download PDF
24. Self-Localization in a Low Cost Bluetooth Environment
- Author
-
Filho, Julio Oliveira, Bunoza, Ana, Sommer, Jürgen, Rosenstiel, Wolfgang, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Doug, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Sandnes, Frode Eika, editor, Zhang, Yan, editor, Rong, Chunming, editor, Yang, Laurence T., editor, and Ma, Jianhua, editor
- Published
- 2008
- Full Text
- View/download PDF
25. The deep learning method applied to the detection and mapping of stone deterioration in open-air sanctuaries of the Hittite period in Anatolia
- Author
-
Ergün Hatır, Mustafa Korkanç, İsmail İnce, and Andreas Schachner
- Subjects
Archeology ,business.industry ,Computer science ,Materials Science (miscellaneous) ,Deep learning ,Pattern recognition ,Conservation ,language.human_language ,Hittite language ,Chemistry (miscellaneous) ,Feature (computer vision) ,language ,Training phase ,Computer vision algorithms ,Artificial intelligence ,business ,General Economics, Econometrics and Finance ,Spectroscopy ,Period (music) ,Open air - Abstract
The detection of deterioration in archeological heritage sites is a very time-consuming task that requires expertise. Furthermore, vision-based approaches can cause errors, considering the complex types of deterioration that develop in different scales and forms in monuments. This problem can be solved effectively owing to computer vision algorithms, commonly used in different areas nowadays. This study aims to develop a model that automatically detects and maps deteriorations (biological colonization, contour scaling, crack, higher plant, impact damage, microkarst, missing part) and restoration interventions using the Mask R-CNN algorithm, which has recently come to the fore with its feature of recognizing small and large-sized objects. To this end, a total of 2460 images of Yazilikaya monuments in the Hattusa archeological site, which is on the UNESCO heritage list, were gathered. In the training phase of the proposed method, it was trained in model 1 to distinguish deposit deterioration commonly observed on the surface of monuments from other anomalies. Other anomalies trained were model 2. In this phase of the models, the average precision values with high accuracy rates ranging from 89.624% to 100% were obtained for the deterioration classes. The developed algorithms were tested on 4 different rock reliefs in Yazilikaya, which were not used in the training phase. In addition, an image of the Eflatunpinar water monument, which is on the UNESCO tentative list, was used to test the model's universality. According to the test results, it was determined that the models could be successfully applied to obtain maps of deterioration and restoration interventions in monuments in different regions.
- Published
- 2021
- Full Text
- View/download PDF
26. Method for Artificial KPI Generation With Realistic Time-Dependent Behaviour
- Author
-
Hao Qiang Luo-Chen, Emil J. Khatib, Carlos S. Alvarez-Merino, and Raquel Barco
- Subjects
Series (mathematics) ,Magnetic reluctance ,Computer science ,computer.software_genre ,Computer Science Applications ,Data modeling ,Modeling and Simulation ,Cellular network ,Training phase ,Data mining ,Performance indicator ,Electrical and Electronic Engineering ,Time series ,computer - Abstract
Machine Learning (ML) is the dominating solution for the implementation of Self-Organizing Networks (SON), which automate mobile network management. However, the data scarcity derived from the reluctance of operators complicates the necessary training phase ML algorithms. In this letter a method to generate artificial Key Performance Indicators (KPIs) time series is proposed considering their time-dependent behaviour. The data is modelled and categorised according to the time of the day and the data models are adapted with statistical copulas to create samples which present interrelation among different KPIs. Finally, results obtained from a real mobile network are presented.
- Published
- 2021
- Full Text
- View/download PDF
27. Ensemble deep transfer learning model for Arabic (Indian) handwritten digit recognition
- Author
-
Ammar Almasri, Wafa’ Za’al Alma’aitah, Nawaf Farhan Funkur Alshdaifat, Rami S. Alkhawaldeh, and Moatsum Alawida
- Subjects
Arabic ,business.industry ,Computer science ,Machine learning ,computer.software_genre ,language.human_language ,Artificial neural network classifier ,Task (project management) ,Artificial Intelligence ,language ,Training phase ,Computational Science and Engineering ,Artificial intelligence ,Digit recognition ,Transfer of learning ,business ,computer ,Noisy data ,Software - Abstract
Recognising handwritten digits or characters is a challenging task due to noisy data that results from different writing styles. Numerous applications essentially motivate to build an effective recognising model for such purposes by utilizing recent intelligent techniques. However, the difficulty emerges when using the Arabic language that suffers from diverse noises; because of the way of writing inherent in connecting characters and digits. Therefore, this work focuses on the Arabic (Indian) digits and propose an ensemble deep transfer learning (EDTL) model that efficaciously detect and recognise these digits. The EDTL model is a combination of two effective pre-trained transfer learning models that consume time and cost complexity in the training phase. The EDTL is trained on large datasets to extract relevant features as input to a fully-connected Artificial Neural Network classifier. The experimental results, using popular datasets, show significant performance obtained by the EDTL model with accuracy reached up to 99.83% in comparison to baseline methods include deep transfer learning models, ensemble deep transfer learning models and state-of-the-art techniques.
- Published
- 2021
- Full Text
- View/download PDF
28. Automated Linking PUBMED Documents with GO Terms Using SVM
- Author
-
Su-Shing Chen and Hyun-Ki Kim
- Subjects
Support vector machine ,ComputingMethodologies_PATTERNRECOGNITION ,Information retrieval ,Computer science ,Gene ontology ,Molecular function ,Genomic data ,Graph (abstract data type) ,Training phase - Abstract
We have developed an automated linking scheme for PUBMED citations with GO terms using SVM (Support Vector Machine), a classification algorithm. The PUBMED database has been essential to life science researchers with over 12 million citations. More recently GO (Gene Ontology) has provided a graph structure for biological process, cellular component, and molecular function of genomic data. By text mining the textual content of PUBMED and associating them with GO terms, we have built up an ontological map for these databases so that users can search PUBMED via GO terms and conversely GO entries via PUBMED classification. Consequently, some interesting and unexpected knowledge may be captured from them for further data analysis and biological experimentation. This paper reports our results on SVM implementation and the need to parallelize for the training phase.
- Published
- 2021
- Full Text
- View/download PDF
29. Optimizing Performance of Automatic Training Phase for Application Performance Prediction in the Grid
- Author
-
Nadeem, Farrukh, Prodan, Radu, Fahringer, Thomas, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Doug, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Perrott, Ronald, editor, Chapman, Barbara M., editor, Subhlok, Jaspal, editor, de Mello, Rodrigo Fernandes, editor, and Yang, Laurence T., editor
- Published
- 2007
- Full Text
- View/download PDF
30. A Neural Network Based on Biological Vision Learning and Its Application on Robot
- Author
-
Gao, Ying, Lu, Xiaodan, Zhang, Liming, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Dough, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Wang, Jun, editor, Liao, Xiao-Feng, editor, and Yi, Zhang, editor
- Published
- 2005
- Full Text
- View/download PDF
31. Deep learning on digital mammography for expert-level diagnosis accuracy in breast cancer detection
- Author
-
Xuran Zhao, Hailiang Li, Zhaoqi Wang, Jinrong Qu, Bailin Yang, Peng Chen, and Zhenzhen Liu
- Subjects
medicine.medical_specialty ,Digital mammography ,Computer Networks and Communications ,Computer science ,education ,CAD ,02 engineering and technology ,Imaging data ,Breast cancer ,0202 electrical engineering, electronic engineering, information engineering ,Media Technology ,medicine ,Medical physics ,business.industry ,Deep learning ,020207 software engineering ,Workload ,medicine.disease ,Cad system ,Hardware and Architecture ,Training phase ,020201 artificial intelligence & image processing ,Artificial intelligence ,business ,Software ,Information Systems - Abstract
Recently, computer-aided diagnosis (CAD) systems powered by deep learning (DL) algorithms have shown excellent performance in the evaluation of digital mammography for breast cancer diagnosis. However, such systems typically require pixel-level annotations by expert radiologists which is prohibitively time-consuming and expensive. Medical institutes would wonder if a high-performance breast cancer CAD system can be trained by exploring their own huge amount of historical imaging data and corresponding diagnosis reports, without additional annotations workload of their radiologists. In this study, we show that a DL classification model trained on historical mammograms with only image-level pathology labels (which can be automatically extracted from medical reports) can achieve surprisingly good diagnostic performance on newly incoming exams compared with experienced radiologists. A DL model called DenseNet was trained and cross-validated with 5979 historical exams acquired before September 2017 with biopsy-verified pathology and tested with 1194 newly obtained cases after that. For both cross-validation and test sets, the ROCs generated by DL predictions were above the ROCs generated by ratings from radiologists. For the suspicious cases which radiologists suggest biopsy (BI-RADS category 4 and 5), the DL model can reject 60% of false biopsies on benign breasts while keeping 95% sensitivity. For the mammograms based on which radiologists were not able to make a diagnosis (BI-RADS 0), the DL model still achieved an AUC score of 79%. Moreover, the model is able to localize lesions on mammograms although such information was not provided in the training phase. Finally, the impact of input image resolution and different DL model architectures on the diagnostic accuracy were also presented and analyzed.
- Published
- 2021
- Full Text
- View/download PDF
32. Improving and Developing the Fog Stability Index for Predicting Fog at Borg El-Arab Airport, Egypt Using WRF Model
- Author
-
Mostafa A. Mohamed, Abdallah Abdeldym, Mostafa Morsy, and Tarek Sayad
- Subjects
Horizontal resolution ,Geophysics ,Meteorology ,Stability index ,Geochemistry and Petrology ,Weather Research and Forecasting Model ,Training phase ,Environmental science ,Regression analysis ,010502 geochemistry & geophysics ,01 natural sciences ,Short duration ,0105 earth and related environmental sciences - Abstract
This study aims to improve and develop local fog stability index (DFSI) as logarithm of horizontal visibility over Borg El-Arab airport, Egypt during the most frequent months of the phenomenon. The results showed that autumn and winter seasons has maximum hourly fog frequency with long duration during the period 1990–2009, so the DFSI regression equations were developed and validated for these seasons. The hourly predictors were obtained from the Weather Research and Forecasting (WRF) model as nest domain with 3 km horizontal resolution along all days of 15 different cases from October to March during the period 1990–2009. Two regression equations were developed for months from October to January (DFSIO-J) and from February March (DFSIF-M). It is found that the developed two regression equations more accurate in training phase than validation phase for fog events. Where, the success percentage for DFSIO-J for fog events reaches to its maximum 100% in January 1997 (training phase) and 91% in November 1996 (validation phase). Whereas, the success percentage for DFSIF-M in February and March are 100% in both two phases except March 89% in the validation phase for fog events. For non-fog events during the two phases, the success percentage for DFSIO-J ranges from 60 to 78%, while it ranges between 64 and 90% for DFSIF-M. Moreover, the general accuracy of DFSIO-J (DFSIF-M) ranges between 61.3 and 77.76% (67.25% and 90.31%) in the training phase and between 63.58 and 78.41% (64.53% and 83.23%) in the validation phase.
- Published
- 2021
- Full Text
- View/download PDF
33. Feasibility of anti-HCV reflex HCV Ag screening strategy in an HCV endemic community
- Author
-
Mei-Yen Chen, Sheng-Nan Lu, Wei-Cheng Huang, Wei-Ming Chen, Tung-Jung Huang, Nien-Tzu Hsu, and Chih-Yi Lee
- Subjects
medicine.medical_specialty ,Medicine (General) ,Silver ,Taiwan ,Test sensitivity ,Hepacivirus ,Gastroenterology ,Sensitivity and Specificity ,03 medical and health sciences ,0302 clinical medicine ,R5-920 ,Positive predicative value ,Internal medicine ,Reflex ,Medicine ,Humans ,HCV Ag ,business.industry ,Anti hiv ,HCV screening ,virus diseases ,General Medicine ,Hepatitis C ,Serum samples ,medicine.disease ,digestive system diseases ,Anti-HCV reflex HCV Ag screening ,030220 oncology & carcinogenesis ,Training phase ,Feasibility Studies ,RNA, Viral ,030211 gastroenterology & hepatology ,Hcv core antigen ,business - Abstract
Background/Purpose: The HCV core antigen (HCV Ag) assay displays high sensitivity and strong correlation with HCV RNA. However, the feasibility of anti-HCV reflex HCV Ag screening in a community-wide setting is rarely discussed. Methods: We performed a two-phase community-based hepatitis C screen in an HCV-prone area of central Taiwan. During the training phase, all participants were test for anti-HCV, HCV Ag and HCV RNA to validate sensitivity, specificity, and accuracy of HCV Ag. During the validation phase, an anti-HCV reflex HCV Ag screen was conducted based on the results of training phase. Outcomes of the study were presented as positive and negative predictive values (PPV and NPV). Results: Of 935 training phase participants, the rate of positive anti-HCV and HCV Ag were 175 (18.7%) and 78 (8.3%), respectively. Test sensitivity, specificity, and accuracy of HCV Ag were 97.1%, 98.6%, and 97.8%, respectively. During validation phase, only anti-HCV-positive serum samples were tested for HCV Ag. Of 1932 participant, 285 (14.8%) were anti-HCV-positive. 133 (46.7%) of the 285 anti-HCV-positive samples were HCV Ag-positive. PPV and NPV were 98.4% and 99.3%, respectively. Across the entire participant sample, a significant linear correlation between HCV Ag and HCV RNA concentration was noted (r2 = 0.93, p-value
- Published
- 2021
34. Count Estimation With a Low-Accuracy Machine Learning Model
- Author
-
Yuichi Sei and Akihiko Ohsuga
- Subjects
0106 biological sciences ,Computer Networks and Communications ,Computer science ,Bayesian probability ,02 engineering and technology ,Machine learning ,computer.software_genre ,010603 evolutionary biology ,01 natural sciences ,0202 electrical engineering, electronic engineering, information engineering ,Estimation ,business.industry ,Confusion matrix ,Object (computer science) ,Computer Science Applications ,Preferred walking speed ,Hardware and Architecture ,Signal Processing ,Deep neural networks ,Training phase ,020201 artificial intelligence & image processing ,Artificial intelligence ,Internet of Things ,business ,computer ,Information Systems - Abstract
Many Internet-of-Things (IoT) systems use machine learning techniques, such as deep neural networks. IoT systems can predict attributes, such as age, sex, car speed, human walking speed, and types of animals, using machine learning techniques. Although the functionality of machine learning is undeniable, the prediction accuracy is not always high. When a machine learning model is used to recognize several objects in an object counting system, the estimated count will have a significant error because of the accumulation of the recognition error of each object. In this study, a count estimation method that uses a confusion matrix generated in the training phase was proposed. The proposed method consists of an iterative Bayesian technique with the confusion matrix for count estimation and mitigating over-iterations technique for reducing estimated errors. The proposed method can be used even for a low-accuracy machine learning model. Experiments with synthetic and real data sets were conducted to demonstrate the functionality of the proposed method. The estimation errors of the proposed method were reduced by 64.3% in average compared to the baseline method in the experiments.
- Published
- 2021
- Full Text
- View/download PDF
35. A influência da fase do treinamento multicomponente no desenvolvimento das valências físicas em atletas universitários de basquetebol / The influence of the multicomponent training phase on the development of physical valences in college basketball athletes
- Author
-
José Camilo Camões, César Rafael Marins Costa, Rafael Bizarelo Ribeiro dos Santos, Raphael da Silva Lau, and Moisés Augusto de Oliveira Borges
- Subjects
Marketing ,Pharmacology ,Organizational Behavior and Human Resource Management ,medicine.medical_specialty ,Basketball ,biology ,Athletes ,Strategy and Management ,Pharmaceutical Science ,biology.organism_classification ,Drug Discovery ,Physical therapy ,medicine ,Training phase ,Psychology - Published
- 2021
- Full Text
- View/download PDF
36. Spam Mail Filtering System Using Semantic Enrichment
- Author
-
Kim, Hyun-Jun, Kim, Heung-Nam, Jung, Jason J., Jo, Geun-Sik, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Dough, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Zhou, Xiaofang, editor, Su, Stanley, editor, Papazoglou, Mike P., editor, Orlowska, Maria Elzbieta, editor, and Jeffery, Keith, editor
- Published
- 2004
- Full Text
- View/download PDF
37. Monitoring the Evolution of Web Usage Patterns
- Author
-
Baron, Steffan, Spiliopoulou, Myra, Hutchison, David, editor, Kanade, Takeo, editor, Kittler, Josef, editor, Kleinberg, Jon M., editor, Mattern, Friedemann, editor, Mitchell, John C., editor, Naor, Moni, editor, Nierstrasz, Oscar, editor, Pandu Rangan, C., editor, Steffen, Bernhard, editor, Sudan, Madhu, editor, Terzopoulos, Demetri, editor, Tygar, Dough, editor, Vardi, Moshe Y., editor, Weikum, Gerhard, editor, Carbonell, Jaime G., editor, Siekmann, Jörg, editor, Berendt, Bettina, editor, Hotho, Andreas, editor, Mladenič, Dunja, editor, van Someren, Maarten, editor, Spiliopoulou, Myra, editor, and Stumme, Gerd, editor
- Published
- 2004
- Full Text
- View/download PDF
38. Audio Metric Learning by Using Siamese Autoencoders for One-Shot Human Fall Detection
- Author
-
Francesco Piazza, Stefano Squartini, Emanuele Principi, Leonardo Gabrielli, and Diego Droghini
- Subjects
Measure (data warehouse) ,One shot ,Control and Optimization ,Artificial neural network ,Computer science ,business.industry ,Pattern recognition ,computer.software_genre ,Expert system ,Computer Science Applications ,Domain (software engineering) ,Computational Mathematics ,Artificial Intelligence ,Metric (mathematics) ,Training phase ,Artificial intelligence ,Fall detection ,business ,computer - Abstract
In the recent years, several supervised and unsupervised approaches to fall detection have been presented in the literature. These are generally based on a corpus of examples of human falls that are, though, hard to collect. For this reason, fall detection algorithms should be designed to gather as much information as possible from the few available data related to the type of events to be detected. The one-shot learning paradigm for expert systems training seems to naturally match these constraints, and this inspired the novel Siamese Neural Network (SNN) architecture for human fall detection proposed in this contribution. Acoustic data are employed as input, and the twin convolutional autoencoders composing the SNN are trained to perform a suitable metric learning in the audio domain and, thus, extract robust features to be used in the final classification stage. A large acoustic dataset has been recorded in three real rooms with different floor types and human falls performed by four volunteers, and then adopted for experiments. Obtained results show that the proposed approach, which only relies on two real human fall events in the training phase, achieves a F $_1$ -Measure of 93.58% during testing, remarkably outperforming the recent supervised and unsupervised state-of-art techniques selected for comparison.
- Published
- 2021
- Full Text
- View/download PDF
39. Comparison of Joint Work During Load Absorption Between Weightlifting Derivatives
- Author
-
Timothy J. Suchomel, Matthew D. Giordanelli, Christopher F. Geiser, and Kristof Kipp
- Subjects
Male ,musculoskeletal diseases ,Knee Joint ,Weight Lifting ,Repetition maximum ,Physical Therapy, Sports Therapy and Rehabilitation ,030204 cardiovascular system & hematology ,03 medical and health sciences ,0302 clinical medicine ,medicine ,Humans ,Orthopedics and Sports Medicine ,Force platform ,Muscle Strength ,Absorption (electromagnetic radiation) ,Exercise ,Joint (geology) ,Mathematics ,business.industry ,Work (physics) ,030229 sport sciences ,General Medicine ,Biomechanical Phenomena ,medicine.anatomical_structure ,Training phase ,Hip Joint ,Ankle ,Nuclear medicine ,business ,Ankle Joint - Abstract
Suchomel, TJ, Giordanelli, MD, Geiser, CF, and Kipp, K. Comparison of joint work during load absorption between weightlifting derivatives. J Strength Cond Res 35(2S): S127-S135, 2021-This study examined the lower-extremity joint-level load absorption characteristics of the hang power clean (HPC) and jump shrug (JS). Eleven Division I male lacrosse players were fitted with 3-dimensional reflective markers and performed 3 repetitions each of the HPC and JS at 30, 50, and 70% of their 1 repetition maximum (1RM) HPC while standing on force plates. Load absorption joint work and duration at the hip, knee, and ankle joints were compared using 3-way repeated-measures mixed analyses of variance. Cohen's d effect sizes were used to provide a measure of practical significance. The JS was characterized by greater load absorption joint work compared with the HPC performed at the hip (p < 0.001, d = 0.84), knee (p < 0.001, d = 1.85), and ankle joints (p < 0.001, d = 1.49). In addition, greater joint work was performed during the JS compared with the HPC performed at 30% (p < 0.001, d = 0.89), 50% (p < 0.001, d = 0.74), and 70% 1RM HPC (p < 0.001, d = 0.66). The JS had a longer loading duration compared with the HPC at the hip (p < 0.001, d = 0.94), knee (p = 0.001, d = 0.89), and ankle joints (p < 0.001, d = 0.99). In addition, the JS had a longer loading duration compared with the HPC performed at 30% (p < 0.001, d = 0.83), 50% (p < 0.001, d = 0.79), and 70% 1RM HPC (p < 0.001, d = 0.85). The JS required greater hip, knee, and ankle joint work on landing compared with the load absorption phase of the HPC, regardless of load. The HPC and JS possess unique load absorption characteristics; however, both exercises should be implemented based on the goals of each training phase.
- Published
- 2021
- Full Text
- View/download PDF
40. Proposing several hybrid PSO-extreme learning machine techniques to predict TBM performance
- Author
-
Danial Jahed Armaghani, Ahmed Mohammed, Edy Tonnizam Mohamad, Bishwajit Roy, Jie Zeng, Deepak Kumar, and Jian Zhou
- Subjects
Schedule ,Coefficient of determination ,business.industry ,Computer science ,0211 other engineering and technologies ,General Engineering ,02 engineering and technology ,Machine learning ,computer.software_genre ,Computer Science Applications ,Acceleration ,020303 mechanical engineering & transports ,0203 mechanical engineering ,Ranking ,Modeling and Simulation ,Tunnel boring machine ,Training phase ,Artificial intelligence ,business ,computer ,Software ,021106 design practice & management ,Extreme learning machine - Abstract
A proper planning schedule for tunnel boring machine (TBM) construction is considered as a necessary and difficult task in tunneling projects. Therefore, prediction of TBM performance with high degree of accuracy is needed to prepare a suitable planning schedule. This study aims to predict the advance rate of TBMs using optimized extreme learning machine (ELM) model with six particles swam optimization (PSO) techniques. Hence, six deterministically adaptive models, including time-varying acceleration (TAC)–PSO–ELM, improved PSO–ELM, Modified PSO–ELM, TAC–MeanPSO–ELM, improved MeanPSO–ELM, and Modified MeanPSO–ELM were developed. A number of performance criteria along with ranking system were used to identify the best model. The results showed that modified MeanPSO–ELM achieved the highest cumulative ranking (56), while the modified PSO–ELM achieved the lowest cumulative ranking (51). For training phase, improved PSO–ELM and TAC–PSO–ELM achieved the highest ranking (30) for each. The TAC–MeanPSO–ELM obtained the lowest ranking in the testing phase (29). Concerning the coefficient of determination (R2), modified PSO–ELM, improved PSO–ELM, TAC–PSO–ELM, and modified MeanPSO–ELM showed a similar behavior and achieved 0.97 for training and 0.96 for testing phases. Two models, including improved MeanPSO–ELM and TAC–MeanPSO–ELM achieved the same R2 of 0.96 for both training and testing phases. The findings of this study suggest that the hybridization of ELM and PSO may result in more accurate results than single ELM model to predict the TBM advance rate.
- Published
- 2021
- Full Text
- View/download PDF
41. A Hybrid Model to Predict Monthly Streamflow Using Neighboring Rivers Annual Flows
- Author
-
Anas Mahmood Al-Juboori
- Subjects
Hydrology ,Water resources ,Hydrogeology ,Streamflow ,Statistical index ,Random tree ,Flow (psychology) ,Training phase ,Environmental science ,Hybrid model ,Water Science and Technology ,Civil and Structural Engineering - Abstract
The issue of predicting monthly streamflow data is an important issue in water resources engineering. In this paper, a hybrid model was proposed to generate monthly streamflow data for a river from the annual streamflow data. To develop the proposed hybrid model, a combination of K-Nearest Neighbor (KNN) and Random Tree (RT) algorithms was used. The hybrid structure was designed to predict the annual flow data for the target river using the annual flow data from neighboring rivers by applying the KNN model, and then generated the monthly flow data for the target river using the predicted annual flow by applying the random tree model. The hybrid model was applied to three rivers in Iraq. The accuracy of the proposed model was tested using two statistical indices, namely, the degree of determination and the efficiency coefficient. The results of the statistical indices indicated a good performance of the proposed hybrid model to generate monthly streamflow using annual streamflow data that the values of degree of determination and efficiency coefficient were greater than 0.91 in the training phase and greater than 0.79 in the validating phase.
- Published
- 2021
- Full Text
- View/download PDF
42. Synthetic datasets for Deep Learning in computer-vision assisted tasks in manufacturing
- Author
-
Nikolaos Nikolakis, Christos Manettas, and Kosmas Alexopoulos
- Subjects
Computer science ,business.industry ,Deep learning ,Machine learning ,computer.software_genre ,Convolutional neural network ,General Earth and Planetary Sciences ,Training phase ,Applications of artificial intelligence ,Artificial intelligence ,business ,computer ,Rotation (mathematics) ,General Environmental Science - Abstract
Artificial Intelligence applications based on Machine Learning methods are widely accepted as promising technologies in manufacturing. Deep Learning (DL) techniques, such as Convolutional Neural Networks (CNN), are successfully used in many computer-vision tasks in manufacturing. These state-of-the-art techniques are requiring large volumes of annotated datasets for training. However, such an approach is expensive, prone to errors and labor as well as time intensive, especially in highly complex and dynamic production environments. Synthetic datasets can be utilized for accelerating the training phase of DL by creating suitable training datasets. This work presents a framework for generating datasets through a chain of simulation tools. The framework is used for generating synthetic images of manufactured parts. States of the parts such as the rotation in different rotation axis need to be recognized by a computer-vision system that assists a manufacturing operation. A number of prior trained CNNs are retrained with the synthetically generated images. The CNNs are tested upon actual images of manufactured parts. The performance of different CNN models is presented, compared and discussed. The results indicate that CNNs trained on synthetically generated datasets may have acceptable performance when used in for assisting tasks in manufacturing.
- Published
- 2021
- Full Text
- View/download PDF
43. The Linear Functional Strategy for Maneuvers Identification Using Elastic Template Matching
- Author
-
Pavlo Tkachenko, Luigi del Re, and Christina Schmid
- Subjects
Dynamic time warping ,Identification (information) ,Current (mathematics) ,Control and Systems Engineering ,Computer science ,Linear form ,Template matching ,Training phase ,Estimator ,Algorithm ,Blossom algorithm - Abstract
In a previous work, we have developed a template (or pattern) matching algorithm for maneuver recognition in highway traffic based on the dynamic time warping alignment strategy. The developed method rely on the model set-ups and tuning parameters, in particular, a template, pre-selected during the training phase. However, the selected template is representing to some extent an “in average” good solution, and the other ones performing less well in general, but maybe better for some particular cases, were rejected. In the current paper we are aiming to introduce a Linear Functional Strategy (LFS) to the community which allows to gain from the variety of possible solutions. This technique belongs to the so-called aggregation approaches, where different models are combined into a final estimator. In our case study, the resulting solution not only allows to avoid pre-selecting of a single template, but also outperforms the best single choice. The latter will be illustrated by numerical experiments on maneuver identification based on a big drone dataset of highway driving.
- Published
- 2021
- Full Text
- View/download PDF
44. Accurate Fine-Grained Layout Analysis for the Historical Tibetan Document Based on the Instance Segmentation
- Author
-
Yuqi Lu, Zhengqi Cai, Penghai Zhao, Weilan Wang, and Guowei Zhang
- Subjects
FOS: Computer and information sciences ,General Computer Science ,Computer science ,Computer Vision and Pattern Recognition (cs.CV) ,Computer Science - Computer Vision and Pattern Recognition ,Inference ,historical Tibetan document images ,computer.software_genre ,Annotation ,text line segmentation ,General Materials Science ,Segmentation ,layout analysis ,business.industry ,General Engineering ,TK1-9971 ,fine-grained layout analysis ,Training phase ,Document analysis and recognition ,Electrical engineering. Electronics. Nuclear engineering ,Stage (hydrology) ,Artificial intelligence ,Line (text file) ,business ,computer ,Character recognition ,Natural language processing ,Sentence - Abstract
Accurate layout analysis without subsequent text-line segmentation remains an ongoing challenge, especially when facing the Kangyur, a kind of historical Tibetan document featuring considerable touching components and mottled background. Aiming at identifying different regions in document images, layout analysis is indispensable for subsequent procedures such as character recognition. However, there was only a little research being carried out to perform line-level layout analysis which failed to deal with the Kangyur. To obtain the optimal results, a fine-grained sub-line level layout analysis approach is presented. Firstly, we introduced an accelerated method to build the dataset which is dynamic and reliable. Secondly, enhancement had been made to the SOLOv2 according to the characteristics of the Kangyur. Then, we fed the enhanced SOLOv2 with the prepared annotation file during the training phase. Once the network is trained, instances of the text line, sentence, and titles can be segmented and identified during the inference stage. The experimental results show that the proposed method delivers a decent 72.7% average precision on our dataset. In general, this preliminary research provides insights into the fine-grained sub-line level layout analysis and testifies the SOLOv2-based approaches. We also believe that the proposed methods can be adopted on other language documents with various layouts., Comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
- Published
- 2021
- Full Text
- View/download PDF
45. IRS-Based TDD Reciprocity Breaking for Pilot Decontamination in Massive MIMO
- Author
-
Hanyu Zhu and Xiliang Luo
- Subjects
Channel reciprocity ,Computer science ,MIMO ,020302 automobile design & engineering ,020206 networking & telecommunications ,02 engineering and technology ,Human decontamination ,Signal-to-noise ratio ,0203 mechanical engineering ,Control and Systems Engineering ,Reciprocity (network science) ,Telecommunications link ,0202 electrical engineering, electronic engineering, information engineering ,Electronic engineering ,Training phase ,Electrical and Electronic Engineering ,Computer Science::Information Theory ,Data transmission - Abstract
In this letter, we propose a novel way to mitigate the notorious pilot contamination in massive multiple-input multiple-output (MIMO) systems by exploiting the intelligent reflecting surfaces (IRSs). Particularly, we intentionally break the uplink (UL) and downlink (DL) channel reciprocity in time-division duplexing (TDD) systems by configuring the phases of the IRSs differently during the UL training phase and the DL data transmission phase. In this way, we show that the pilot contamination is mitigated and can be eliminated in some cases. Meanwhile, we analyze the signal-to-interference-plus-noise ratio (SINR) performance due to the non-reciprocal UL and DL channels. We further show the asymptotic optimality of our proposed decontamination scheme. The analytical results are also corroborated by numerical simulations.
- Published
- 2021
- Full Text
- View/download PDF
46. Wi-Fi location fingerprinting using an intelligent checkpoint sequence.
- Author
-
Retscher, Günther and Hofer, Hannes
- Subjects
- *
WIRELESS Internet , *INTELLIGENT sensors , *GLOBAL Positioning System , *COMPUTERS in geology , *GRAPH theory - Abstract
For Wi-Fi positioning location fingerprinting is very common but has the disadvantage that it is very labour consuming for the establishment of a database (DB) with received signal strength (RSS) scans measured on a large number of known reference points (RPs). To overcome this drawback a novel approach is developed which uses a logical sequence of intelligent checkpoints (iCPs) instead of RPs distributed in a regular grid. The iCPs are the selected RPs which have to be passed along the way for navigation from a start point A to the destination B. They are twofold intelligent because of the fact that they depend on their meaningful selection and because of their logical sequence in their correct order. Thus, always the following iCP is known due to a vector graph allocation in the DB and only a small limited number of iCPs needs to be tested when matching the current RSS scans. This reduces the required processing time significantly. It is proven that the iCP approach achieves a higher success rate than conventional approaches. In average correct matching results of 90.0% were achieved using a joint DB including RSS scans of all employed smartphones. An even higher success rate is achieved if the same mobile device is used in both the training and positioning phase. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
47. 髙级拳击运动员在不同海拔训练时身体运动功能状态变化的特点.
- Author
-
科尔任涅甫斯基, A. H., 库尔古佐夫, R. B., 克连达尔, B. A., and 马莫契金, O. M.
- Abstract
Copyright of Journal of Capital Institute of Physical Education is the property of Shoudu Tiyu Xueyuan and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2019
- Full Text
- View/download PDF
48. A Comparative Study on the Performance of Classification Algorithms for Effective Diagnosis of Liver Diseases
- Author
-
Bihter Daş
- Subjects
Liver diseases,early diagnosis,data classifiers,data mining ,Bilgisayar Bilimleri, Yapay Zeka ,Artificial neural network ,Computer science ,business.industry ,Decision tree ,Bilgisayar Bilimleri, Yazılım Mühendisliği ,General Medicine ,Computer Science, Software Engineering ,Machine learning ,computer.software_genre ,medicine.disease ,Support vector machine ,Tree (data structure) ,Statistical classification ,Liver disease ,Computer Science, Artifical Intelligence ,medicine ,Classification methods ,Training phase ,Artificial intelligence ,business ,computer - Abstract
In recent years, different approaches and methods have been proposed to diagnose various diseases accurately. Since there are a variety of liver diseases, till late-stage liver disease and liver failure occur the symptoms tend to be specific for that illness. Therefore, early diagnosis can play a key role in preventing deaths from liver diseases. In this study, we compare the accuracy of different classification methods supported by the SAS software suite, such as Neural Network, Auto Neural, High Performance (HP) SVM, HP Forest, HP Tree (Decision Tree), and HP Neural for the diagnosis of liver diseases. In this study, the Indian Liver Patient Dataset (ILPD) provided by the University of California, Irvine (UCI) repository is used. Experimental results show that based on the metrics of our study, in the training phase while HP Forest achieves the highest accuracy rate, HP SVM and HP Tree do the lowest accuracy rates. However, in the validation phase, Neural Network achieves the highest accuracy rate and HP Forest does the lowest accuracy rate. Our experimental results may be useful for both researchers and practitioners working in related fields.
- Published
- 2020
- Full Text
- View/download PDF
49. Image enhancement using convolutional neural network to identify similar patterns
- Author
-
Jia-An Lin, Ruei-Han Chen, Ling-Ling Wang, and Ching-Ta Lu
- Subjects
business.industry ,Computer science ,Pattern recognition ,Image processing ,Image enhancement ,Impulse noise ,Convolutional neural network ,Signal Processing ,Training phase ,Computer Vision and Pattern Recognition ,Artificial intelligence ,Electrical and Electronic Engineering ,Image denoising ,business ,Software - Published
- 2020
- Full Text
- View/download PDF
50. Extending the weightless WiSARD classifier for regression
- Author
-
Priscila M. V. Lima, Gabriel P. Guarisa, Aluizio Lima Filho, Lucca M. Felix, Luiz Felipe Ramalho de Oliveira, Felipe M. G. França, and Leopoldo Lusquino Filho
- Subjects
0209 industrial biotechnology ,Artificial neural network ,business.industry ,Cognitive Neuroscience ,Mean absolute error ,Pattern recognition ,02 engineering and technology ,Regression ,Computer Science Applications ,020901 industrial engineering & automation ,Artificial Intelligence ,Weightless ,0202 electrical engineering, electronic engineering, information engineering ,Palm oil ,Training phase ,020201 artificial intelligence & image processing ,Artificial intelligence ,business ,Classifier (UML) ,Mathematics - Abstract
This paper explores two new weightless neural network models, Regression WiSARD and ClusRegression WiSARD, in the challenging task of predicting the total palm oil production of a set of 28 (twenty eight) differently located sites under different climate and soil profiles. Both models were derived from Kolcz and Allinson’s n-Tuple Regression weightless neural model and obtained mean absolute error (MAE) rates of 0.09097 and 0.09173, respectively. Such results are very competitive with the state-of-the-art (0.07983), whilst being four orders of magnitude faster during the training phase. Additionally the models have been tested on three classic regression datasets, also presenting competitive performance with respect to other models often used in this type of task.
- Published
- 2020
- Full Text
- View/download PDF
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.