Your browser doesn't support javascript.
loading
Mostrar: 20 | 50 | 100
Resultados 1 - 20 de 38
Filtrar
1.
IEEE Trans Cybern ; PP2021 Apr 21.
Artículo en Inglés | MEDLINE | ID: mdl-33882011

RESUMEN

Given m d-dimensional responsors and n d-dimensional predictors, sparse regression finds at most k predictors for each responsor for linear approximation, 1≤ k ≤ d-1. The key problem in sparse regression is subset selection, which usually suffers from high computational cost. In recent years, many improved approximate methods of subset selection have been published. However, less attention has been paid to the nonapproximate method of subset selection, which is very necessary for many questions in data analysis. Here, we consider sparse regression from the view of correlation and propose the formula of conditional uncorrelation. Then, an efficient nonapproximate method of subset selection is proposed in which we do not need to calculate any coefficients in the regression equation for candidate predictors. By the proposed method, the computational complexity is reduced from O([1/6]k³+(m+1)k²+mkd) to O([1/6]k³+[1/2](m+1)k²) for each candidate subset in sparse regression. Because the dimension d is generally the number of observations or experiments and large enough, the proposed method can greatly improve the efficiency of nonapproximate subset selection. We also apply the proposed method in real scenarios of dental age assessment and sparse coding to validate the efficiency of the proposed method.

2.
Inf Sci (N Y) ; 2021 Apr 08.
Artículo en Inglés | MEDLINE | ID: mdl-33846657

RESUMEN

Early warning is a vital component of emergency repsonse systems for infectious diseases. However, most early warning systems are centralized and isolated, thus there are potential risks of single evidence bias and decision-making errors. In this paper, we tackle this issue via proposing a novel framework of collaborative early warning for COVID-19 based on blockchain and smart contracts, aiming to crowdsource early warning tasks to distributed channels including medical institutions, social organinzations, and even individuals. Our framework supports two surveillance modes, namely, medical federation surveillance based on federated learning and social collaboration surveillance based on the learning markets approach, and fuses their monitoring results on emerging cases to alert. By using our framework, medical institutions are expected to obtain better federated surveillance models with privacy protection, and social participants without mutual trusts can also share verified surveillance resources such as data and models, and fuse their surveillance solutions. We implemented our proposed framework based on the Ethereum and IPFS platforms. Experimental results show that our framework has advantages of decentralized decision-making, fairness, auditability, and universality, and also has potential guidance and reference value for the early warning and prevention of unknown infectious diseases.

3.
Comput Methods Programs Biomed ; 202: 106019, 2021 Apr.
Artículo en Inglés | MEDLINE | ID: mdl-33640650

RESUMEN

BACKGROUND AND OBJECTIVE: In medical imaging, the scarcity of labeled lesion data has hindered the application of many deep learning algorithms. To overcome this problem, the simulation of diverse lesions in medical images is proposed. However, synthesizing labeled mass images in mammograms is still challenging due to the lack of consistent patterns in shape, margin, and contextual information. Therefore, we aim to generate various labeled medical images based on contextual information in mammograms. METHODS: In this paper, we propose a novel approach based on GANs to generate various mass images and then perform contextual infilling by inserting the synthetic lesions into healthy screening mammograms. Through incorporating features of both realistic mass images and corresponding masks into the adversarial learning scheme, the generator can not only learn the distribution of the real mass images but also capture the matching shape, margin and context information. RESULTS: To demonstrate the effectiveness of our proposed method, we conduct experiments on publicly available mammogram database of DDSM and a private database provided by Nanfang Hospital in China. Qualitative and quantitative evaluations validate the effectiveness of our approach. Additionally, through the data augmentation by image generation of the proposed method, an improvement of 5.03% in detection rate can be achieved over the same model trained on original real lesion images. CONCLUSIONS: The results show that the data augmentation based on our method increases the diversity of dataset. Our method can be viewed as one of the first steps toward generating labeled breast mass images for precise detection and can be extended in other medical imaging domains to solve similar problems.

4.
Artículo en Inglés | MEDLINE | ID: mdl-33600323

RESUMEN

Image ordinal estimation is to predict the ordinal label of a given image, which can be categorized as an ordinal regression (OR) problem. Recent methods formulate an OR problem as a series of binary classification problems. Such methods cannot ensure that the global ordinal relationship is preserved since the relationships among different binary classifiers are neglected. We propose a novel OR approach, termed convolutional OR forest (CORF), for image ordinal estimation, which can integrate OR and differentiable decision trees with a convolutional neural network for obtaining precise and stable global ordinal relationships. The advantages of the proposed CORF are twofold. First, instead of learning a series of binary classifiers independently, the proposed method aims at learning an ordinal distribution for OR by optimizing those binary classifiers simultaneously. Second, the differentiable decision trees in the proposed CORF can be trained together with the ordinal distribution in an end-to-end manner. The effectiveness of the proposed CORF is verified on two image ordinal estimation tasks, i.e., facial age estimation and image esthetic assessment, showing significant improvements and better stability over the state-of-the-art OR methods.

5.
Accid Anal Prev ; 150: 105937, 2021 Feb.
Artículo en Inglés | MEDLINE | ID: mdl-33338914

RESUMEN

The autonomous vehicle is regarded as a promising technology with the potential to reshape mobility and solve many traffic issues, such as accessibility, efficiency, convenience, and especially safety. Many previous studies on driving strategies mainly focused on the low-level detailed driving behaviors or specific traffic scenarios but lacked the high-level driving strategy studies. Though researchers showed increasing interest in driving strategies, there still has no comprehensive answer on how to proactively implement safe driving. After analyzing several representative driving strategies, we propose three characteristic dimensions that are important to measure driving strategies: preferred objective, risk appetite, and collaborative manner. According to these three characteristic dimensions, we categorize existing driving strategies of autonomous vehicles into four kinds: defensive driving strategies, competitive driving strategies, negotiated driving strategies, and cooperative driving strategies. This paper provides a timely comparative review of these four strategies and highlights the possible directions for improving the high-level driving strategy design.

6.
IEEE Trans Cybern ; PP2020 Oct 23.
Artículo en Inglés | MEDLINE | ID: mdl-33095725

RESUMEN

Multiagent reinforcement learning (MARL) has recently attracted considerable attention from both academics and practitioners. Core issues, e.g., the curse of dimensionality due to the exponential growth of agent interactions and nonstationary environments due to simultaneous learning, hinder the large-scale proliferation of MARL. These problems deteriorate with an increased number of agents. To address these challenges, we propose an adversarial collaborative learning method in a mixed cooperative-competitive environment, exploiting friend-or-foe Q-learning and mean-field theory. We first treat neighbors of agent i as two coalitions (i's friend and opponent coalition, respectively), and convert the Markov game into a two-player zero-sum game with an extended action set. By exploiting mean-field theory, this new game simplifies the interactions as those between a single agent and the mean effects of friends and opponents. A neural network is employed to learn the optimal mean effects of these two coalitions, which are trained via adversarial max and min steps. In the max step, with fixed policies of opponents, we optimize the friends' mean action to maximize their rewards. In the min step, the mean action of opponents is trained to minimize the friends' rewards when the policies of friends are frozen. These two steps are proved to converge to a Nash equilibrium. Then, another neural network is applied to learn the best response of each agent toward the mean effects. Finally, the adversarial max and min steps can jointly optimize the two networks. Experiments on two platforms demonstrate the learning effectiveness and strength of our approach, especially with many agents.

7.
IEEE Trans Cybern ; PP2020 Jun 01.
Artículo en Inglés | MEDLINE | ID: mdl-32479409

RESUMEN

Population synthesis is the foundation of the agent-based social simulation. Current approaches mostly consider basic population and households, rather than other social organizations. This article starts with a theoretical analysis of the iterative proportional updating (IPU) algorithm, a representative method in this field, and then gives an extension to consider more social organization types. The IPU method, for the first time, proves to be unable to converge to an optimal population distribution that simultaneously satisfies the constraints from individual and household levels. It is further improved to a bilevel optimization, which can solve such a problem and include more than one type of social organization. Numerical simulations, as well as population synthesis using actual Chinese nationwide census data, support our theoretical conclusions and indicate that our proposed bilevel optimization can both synthesize more social organization types and get more accurate results.

8.
Technol Health Care ; 28(S1): 131-150, 2020.
Artículo en Inglés | MEDLINE | ID: mdl-32364146

RESUMEN

BACKGROUND: Endoscopic endonasal transsphenoidal pituitary surgery is usually difficult and risky. With limited sources of cadaveric skulls, traditional methods of using virtual images to study the surgery are difficult for neurosurgeons and students because the surgery requires spatial imagination and good understanding of the patient's conditions as well as practical experience. The three-dimensional (3D) printing technique has played an important role in clinical medicine due to its advantages of low cost, high-efficiency and customization. OBJECTIVE: CT images are used as the source data of 3D printing. The data obtained directly from the CT machine has limited accuracy, which cannot be printed without processing. Some commercial platforms can help build an accurate model but the cost and customization are not satisfactory. In this situation, a tactile, precise and low-cost 3D model is highly desirable. METHODS: Five kinds of computer software are used in the manufacturing of medical 3D models and the processing procedure is easy to understand and operate. RESULTS: This study proposes a practical and cost-effective method to obtain the corrected digital model and produce the 3D printed skull with complete structures of nasal cavity, sellar region and different levels of pituitary tumors. The model is used for the endoscopic endonasal transsphenoidal pituitary surgery preparation. CONCLUSION: The 3D printed medical model can directly help neurosurgeons and medical students to practice their surgery skills on both general and special cases with customized structures and different levels of tumors.

9.
IEEE Trans Neural Syst Rehabil Eng ; 28(2): 488-497, 2020 02.
Artículo en Inglés | MEDLINE | ID: mdl-31902766

RESUMEN

Assessing facial nerve function from visible facial signs such as resting asymmetry and symmetry of voluntary movement is an important means in clinical practice. By using image processing, computer vision and machine learning techniques, replacing the clinician with a machine to do assessment from ubiquitous visual face capture is progressing more closely to reality. This approach can do assessment in a purely automated manner, hence opens a promising direction for future development in this field. Many studies gathered around this interesting topic with a variety of solutions proposed in recent years. However, to date, none of these solutions have gained a widespread clinical use. This study provides a comprehensive review of the most relevant and representative studies in automated facial nerve function assessment from visual face capture, aiming at identifying the principal challenges in this field and thus indicating directions for future work.

10.
IEEE Trans Cybern ; PP2020 Jan 23.
Artículo en Inglés | MEDLINE | ID: mdl-31985449

RESUMEN

In this article, we propose a key secret-sharing technology based on generative adversarial networks (GANs) to address three major problems in the blockchain: 1) low security; 2) hard recovery of lost keys; and 3) low communication efficiency. In our scheme, the proposed network plays the role of a dealer and treats the secret-sharing process as a classification issue. The key idea is to view the secret as an image during the secret-sharing process. If the user's private key is text, we can covert the key text into an image called the original image. Specifically, we first divide the original image into original subimages by the image segmentation. Next, we encode each original subimage by DNA coding. Finally, we train the proposed network to find the key secret-sharing results. Our proposed scheme is not only a significant extension of the GANs but also a new direction for the key secret-sharing technology. The simulation results show that the scheme is secure, and both flexible and efficient in communication.

11.
IEEE Trans Cybern ; PP2020 Jan 16.
Artículo en Inglés | MEDLINE | ID: mdl-31945004

RESUMEN

Agent-based simulation is a useful approach for the analysis of dynamic population evolution. In this field, the existing models mostly treat the migration behavior as a result of utility maximization, which partially ignores the endogenous mechanisms of human decision making. To simulate such a process, this article proposes a new cognitive architecture called the two-layered integrated decision cycle (TiDEC) which characterizes the individual's decision-making process. Different from the previous ones, the new hybrid architecture incorporates deep neural networks for its perception and implicit knowledge learning. The proposed model is applied in China and U.S. population evolution. To the best of our knowledge, this is the first time that the cognitive computation is used in such a field. Computational experiments using the actual census data indicate that the cognitive model, compared with the traditional utility maximization methods, cannot only reconstruct the historical demographic features but also achieve better prediction of future evolutionary dynamics.

12.
IEEE Trans Neural Netw Learn Syst ; 31(11): 4649-4659, 2020 Nov.
Artículo en Inglés | MEDLINE | ID: mdl-31899442

RESUMEN

Machine learning, especially deep neural networks, has developed rapidly in fields, including computer vision, speech recognition, and reinforcement learning. Although minibatch stochastic gradient descent (SGD) is one of the most popular stochastic optimization methods for training deep networks, it shows a slow convergence rate due to the large noise in the gradient approximation. In this article, we attempt to remedy this problem by building a more efficient batch selection method based on typicality sampling, which reduces the error of gradient estimation in conventional minibatch SGD. We analyze the convergence rate of the resulting typical batch SGD algorithm and compare the convergence properties between the minibatch SGD and the algorithm. Experimental results demonstrate that our batch selection scheme works well and more complex minibatch SGD variants can benefit from the proposed batch selection strategy.

13.
IEEE Trans Image Process ; 29(1): 2078-2093, 2020.
Artículo en Inglés | MEDLINE | ID: mdl-31647437

RESUMEN

We propose Mask SSD, an efficient and effective approach to address the challenging instance segmentation task. Based on a single-shot detector, Mask SSD detects all instances in an image and marks the pixels that belong to each instance. It consists of a detection subnetwork that predicts object categories and bounding box locations, and an instance-level segmentation subnetwork that generates the foreground mask for each instance. In the detection subnetwork, multi-scale and feedback features from different layers are used to better represent objects of various sizes and provide high-level semantic information. Then, we adopt an assistant classification network to guide per-class score prediction, which consists of objectness prior and category likelihood. The instance-level segmentation subnetwork outputs pixel-wise segmentation for each detection while providing the multi-scale and feedback features from different layers as input. These two subnetworks are jointly optimized by a multi-task loss function, which renders Mask SSD direct prediction on detection and segmentation results. We conduct extensive experiments on PASCAL VOC, SBD, and MS COCO datasets to evaluate the performance of Mask SSD. Experimental results verify that as compared with state-of-the-art approaches, our proposed method has a comparable precision with less speed overhead.

14.
IEEE Trans Neural Netw Learn Syst ; 31(3): 801-812, 2020 Mar.
Artículo en Inglés | MEDLINE | ID: mdl-31071054

RESUMEN

As one of the efficient approaches to deal with big data, divide-and-conquer distributed algorithms, such as the distributed kernel regression, bootstrap, structured perception training algorithms, and so on, are proposed and broadly used in learning systems. Some learning theories have been built to analyze the feasibility, approximation, and convergence bounds of these distributed learning algorithms. However, less work has been studied on the stability of these distributed learning algorithms. In this paper, we discuss the generalization bounds of distributed learning algorithms from the view of algorithmic stability. First, we introduce a definition of uniform distributed stability for distributed algorithms and study the distributed algorithms' generalization risk bounds. Then, we analyze the stability properties and generalization risk bounds of a kind of regularization-based distributed algorithms. Two generalization distributed risks obtained show that the generalization distributed risk bounds for the difference between their generalization distributed and empirical distributed/leave-one-computer-out risks are closely related to the size of samples n and the amount of working computers m as O(m/n1/2) . Furthermore, the results in this paper indicate that, for a good generalization regularized distributed kernel algorithm, the regularization parameter λ should be adjusted with the change of the term m/n1/2 . These theoretic discoveries provide the useful guidance when deploying the distributed algorithms on practical big data platforms. We explore our theoretic analyses through two simulation experiments. Finally, we discuss some problems about the sufficient amount of working computers, nonequivalence, and generalization for distributed learning. We show that the rules for the computation on one single computer may not always hold for distributed learning.

15.
Comput Methods Programs Biomed ; 180: 105012, 2019 Oct.
Artículo en Inglés | MEDLINE | ID: mdl-31421601

RESUMEN

BACKGROUND AND OBJECTIVE: Simulation of diverse lesions in images is proposed and applied to overcome the scarcity of labeled data, which has hindered the application of deep learning in medical imaging. However, most of current studies focus on generating samples with class labels for classification and detection rather than segmentation, because generating images with precise masks remains a challenge. Therefore, we aim to generate realistic medical images with precise masks for improving lesion segmentation in mammagrams. METHODS: In this paper, we propose a new framework for improving X-ray breast mass segmentation performance aided by generated adversarial lesion images with precise masks. Firstly, we introduce a conditional generative adversarial network (cGAN) to learn the distribution of real mass images as well as a mapping between images and corresponding segmentation masks. Subsequently, a number of lesion images are generated from various binary input masks using the generator in the trained cGAN. Then the generated adversarial samples are concatenated with original samples to produce a dataset with increased diversity. Furthermore, we introduce an improved U-net and train it on the previous augmented dataset for breast mass segmentation. RESULTS: To demonstrate the effectiveness of our proposed method, we conduct experiments on publicly available mammogram database of INbreast and a private database provided by Nanfang Hospital in China. Experimental results show that an improvement up to 7% in Jaccard index can be achieved over the same model trained on original real lesion images. CONCLUSIONS: Our proposed method can be viewed as one of the first steps toward generating realistic X-ray breast mass images with masks for precise segmentation.


Asunto(s)
Aprendizaje Profundo , Procesamiento de Imagen Asistido por Computador/métodos , Mamografía , Rayos X , China , Femenino , Humanos
16.
Comput Assist Surg (Abingdon) ; 24(sup1): 121-130, 2019 10.
Artículo en Inglés | MEDLINE | ID: mdl-31012745

RESUMEN

In general, the 3 D printed medical models are made based on virtual digital models obtained from machines such as the computed tomography scanner. However, due to the limited accuracy of CT scanning technology, which is usually 1 millimeter, there are differences between scanned results and the real structure. Besides, the collected data can hardly be printed directly because of some errors in the model. In this paper, we present a general and efficient procedure to process the digital skull data to make the printed structures meet the requirements of anatomy education, which combines the use of five 3 D manipulation tools and the procedure can be finished within 6 hours. Then the model is printed and compared with the cadaveric skull from frontal, left, right and anterior views respectively. The printed model can describe the correct structure and details of the skull clearly, which can be considered as a good alternative to the cadaveric skull. The manipulation procedure presented in this study is an easily available and cost-effective way to obtain a printed skull model from the original CT data, which has a considerable economic and social benefit for the medical education. The steps of the data processing can be performed easily. The cost for the 3 D printed model is also low. Outcomes of this study can be applied widely in processing skull data.


Asunto(s)
Anatomía/educación , Modelos Anatómicos , Impresión Tridimensional , Cráneo/anatomía & histología , Educación de Pregrado en Medicina , Humanos , Cráneo/diagnóstico por imagen , Tomografía Computarizada por Rayos X , Interfaz Usuario-Computador
17.
IEEE Trans Cybern ; 49(11): 4042-4050, 2019 Nov.
Artículo en Inglés | MEDLINE | ID: mdl-30843813

RESUMEN

As a typical cyber-physical system, 3D printing has developed very fast in recent years. There is a strong demand for mass customization, such as printing dental crowns. However, the accuracy of the 3D printed objects is low compared with traditional methods. The main reason is that the model to be printed is arbitrary and usually the quantity is small. The deformation is affected by the shape of the object and there is a lack of a universal method for the error compensation. It is neither easy nor economical to perform the compensation manually. In this paper, we present a framework for the automatic error compensation. We obtain the shape by technologies such as 3D scanning. And we use the "3D deep learning" method to train a deep neural network. For a specific task, such as dental crown printing, the network can learn the function of deformation when a large amount of data is used for training. To the best of our knowledge, this is the first application of the deep neural network to the error compensation in 3D printing. And we propose the "inverse function network" to compensate for the error. We use four types of deformations of the dental crowns to verify the performance of the neural network: 1) translation; 2) scaling up; 3) scaling down; and 4) rotation. The convolutional AutoEncoder structure is employed for the end-to-end learning. The experiments show that the network can predict and compensate for the error well. By introducing the new method, we can improve the accuracy with little need for increasing the hardware cost.

18.
Sci Robot ; 4(28)2019 Mar 27.
Artículo en Inglés | MEDLINE | ID: mdl-33137752

RESUMEN

A self-driven closed-loop parallel testing system implements more challenging tests to accelerate evaluation and development of autonomous vehicles.

19.
IEEE Trans Cybern ; 49(10): 3618-3626, 2019 Oct.
Artículo en Inglés | MEDLINE | ID: mdl-30571655

RESUMEN

To give a high-level summary to current approaches for implementing artificial intelligence (AI), we explain the key commonalities and major differences between Turing's approach and Wiener's approach in this perspective. Especially, the problems, successful achievements, limitations, and future research directions of existing approaches that follow Weiner's ideas are addressed, respectively, aiming to provide readers with a good start point and a roadmap. Some other related topics, for example, the role of human experts in developing AI, are also discussed to seek potential solutions for some existing difficulties.

20.
IEEE Trans Cybern ; 48(12): 3280-3290, 2018 Dec.
Artículo en Inglés | MEDLINE | ID: mdl-30072355

RESUMEN

Building autonomous systems that achieve human level intelligence is one of the primary objectives in artificial intelligence (AI). It requires the study of a wide range of functions robustly across different phases of human cognition. This paper presents a review of agent cognitive architectures in the past 20 year's AI research. Different from software structures and simulation environments, most of the architectures concerned are established from mathematics and philosophy. They are categorized according to their knowledge processing patterns-symbolic, emergent or hybrid. All the relevant literature can be accessed publicly, particularly through the Internet. Available websites are also summarized for further reference.


Asunto(s)
Inteligencia Artificial , Modelos Neurológicos , Redes Neurales de la Computación , Cognición/fisiología , Humanos , Memoria/fisiología
SELECCIÓN DE REFERENCIAS
DETALLE DE LA BÚSQUEDA
...