Use Of Explicit Information Research Articles

Training a dynamically configurable classifier with deep Q-learning

Subject of study. We studied dynamic networks capable of performing calculations from input data. Aim. We studied whether deep Q-learning can be used for the construction of dynamic computer vision networks. Methods. In modern dynamically configurable systems, image analysis is typically performed using a policy gradient algorithm. We propose a method for hybrid Q-learning by an image classification agent taking into account limitations on available computer resources. We train the agent to recognize images using a set of pretrained classifiers, and the resulting dynamically configurable system is capable of constructing a computational graph that takes into account the limitations on the number of operations with a trajectory that corresponds to the maximum expected accuracy. The agent only receives an award when the image is correctly recognized within a limit on the number of actions that can be taken by the agent. Experiments were performed using the CIFAR-10 image database and a set of six external classifiers that the agent was trained to control. The experiments performed showed that the standard deep learning method using action values (Deep Q-Network) does not permit the agent to learn strategies that are better than random ones in terms of recognition accuracy. We therefore propose a Q-least-action classifier that approximates the desired classifier selection function by reinforcement learning and the label prediction function by supervised learning. Main results. The trained agent exceeded the recognition accuracy of random strategies (reduces the error by 9.65%). We show that such an agent can make explicit use of information from several classifiers since the accuracy increases when the number of permitted actions increases. Practical significance. Our research shows that the deep Q-learning method is capable of extracting information from sparse responses by classifiers as well as a least-action classifier trained by the policy-gradient method. In addition, the method proposed herein did not require the development of special loss functions.

Common spatial patterns combined with phase synchronization information for classification of EEG signals

The common spatial patterns (CSP) approach is a classical and representative technique of optimizing spatial filters of electroencephalogram (EEG) signals in the community of brain computer interfaces (BCI). It, however, utilizes only amplitude information of the EEG signals. The phase information of the multi-channel EEG series, on the other hand, plays an important role in characterizing brain activities. In this paper, we consider enhancing the classification performance of CSP by making explicit use of information of the phase synchronization. An index, termed as rank-weighted phase lag index (rWPLI), is introduced to qualify the intrinsic phase synchronization. The rWPLI features are then incorporated into the CSP framework via three ways of feature combinations. The classification experiments on real EEG data sets of BCI competitions show the effectiveness of the proposed framework.

Modelos Comunicativos y Estrategias de Afrontamiento en la Formación Artística de Maestros de Educación Infantil

Education must address not only the transmission of content, but also the development of cognitive skills not directly obtainable through the acquisition and use of explicit information, as are, for instance coping strategies. The objective of this work is to evaluate a crucial aspect of emotional rationality, i.e., coping strategies, as exemplified by the students of the Early Childhood Education degree at UB, with the aim of proposing improvements in the contents of the degree through the incorporation of arts education activities in the curriculum. In order to do it, we have completed a descriptive correlational study of 98 students by means of the application CRI-A (Moos, 2010) and its subsequent statistical analysis by means of SPSS. The results show a low use of the mentioned strategies (seeking guidance and support, and acceptance and resignation) especially among younger students. On the other hand, a further objective of this work is to argue in favor of a certain conception of communication and education: communicating is not just transmitting information, and the same applies in the field of education.

The selection method for hyper‐parameters of support vector classification by adaptive chaotic cultural algorithm

PurposeThe purpose of this paper is to find a novel optimization selection method for hyper‐parameter of support vector classification (SVC), responsible for the classification of datasets from the UCI machine learning database repository.Design/methodology/approachA novel two‐stage optimization selection method for hyper‐parameters is proposed. It makes use of explicit information derived from issues and implicit knowledge extracted from the evolution process so as to improve the performance of classifier. In the first stage, the search extent of each hyper‐parameter is determined according to the requirements of issues. In the second stage, optimal hyper‐parameters are obtained by adaptive chaotic culture algorithm in the above search extent. Adaptive chaotic cultural algorithm uses implicit knowledge extracted from the evolution process to control mutation scale of chaotic mutation operator. This algorithm can ensure the diversity of population and exploitation in the latter evolution.FindingsThe rationality of the above optimization selection method is proved by the binary classification problem. Final confirmation of this approach is the classification results compared with other methods.Originality/valueThis optimization selection method can effectively avoid premature convergence and lead to better computation stability and precision. It is not related on the structure of functions. SVC model corresponding to optimal hyper‐parameters by this method has better generalization.

How (Not) to Lie with Spatial Statistics

How (Not) to Lie with Spatial Statistics

QFD: an interactive algorithm for the prioritization of product’s technical design characteristics

The paper is concerned with the problem of the “prioritization” of technical design characteristics of a product. An interactive algorithm has been developed with the aim to better support the engineering design process by means of quality function deployment (QFD). The algorithm tries to soften customer approach to QFD in those situations in which customers are not able to give a “significant” evaluation of the relative importance of their requirements. The method allows determining a ranking order of design characteristics without the artificial conversion of symbols contained in the relationship matrix, and without the use of explicit information concerning the relative degree of importance of customer requirements. A simple numerical application is also provided.

Clinical Decision‐Making for Dental Caries Management

Scientific information on diagnosis, prevention, and management of dental caries and associated indicators of risk continues to increase rapidly. Patients vary in clinically important ways, however, and uncertainty affects our understanding of risk; diagnostic and prognostic information; efficacy and effectiveness of many preventive, diagnostic, and treatment alternatives; and outcomes associated with clinical strategies. Consequently, challenges abound for clinicians to identify, evaluate, and incorporate new information, patient preferences, and uncertainties into clinical practice. Clinical decision-making--an analytical approach that makes explicit use of information to quantify probabilities and outcomes to analyze decisions under conditions of uncertainty--can provide a framework to analyze the impact of uncertainty of clinical information. Diagnostic, effectiveness, and outcome information is quantified and combined in an explicit way to serve as a tool for clinicians, not as a replacement for clinical judgment or experience. Such an approach has the potential to improve clinical practice and help dentists do their jobs better by structuring the decision problem and assessing probabilities and utilities. Clinical decision-making also helps dentists communicate with each other by identifying clinical controversies, thereby characterizing how and why disagreements may arise and what additional data may be needed to address a clinical question.

Bayesian Forecasting of the Human Immunodeficiency Virus Epidemic in Scotland

The most commonly used method for acquired immune deficiency syndrome (AIDS) forecasting (backprojection) assumes that the incubation distribution of the time from infection with human immunodeficiency virus (HIV) to AIDS is known, and makes no explicit use of information on the number of individuals who have been infected with HIV. Now that we are further into the epidemic our predictions are more sensitive to assumptions about the incubation distribution which may now be influenced by pre-AIDS treatment. At the same time we have gained more knowledge about the HIV infection curve in Scotland as a result of retrospective testing of stored samples. In addition to AIDS case reports data are collected in Scotland on CD4 counts for HIV positive patients under medical care, which give incidence data on CD200 cases (two consecutive CD4 counts below 200 or an AIDS diagnosis). We have developed Bayesian methodology to make use of our partial knowledge of both the incubation distribution and the infection curve to make short-term forecasts of AIDS and CD200 cases. The range of forecasts from this approach includes the uncertainty in our specification of both the incubation distribution and the infection curve.