Patent classification systems are used to help scrutinize patent applications for possible violations of the novelty and non-obviousness/ inventive steps of a patentability test. There are several different patent cla...
详细信息
In this research will show a method for sound-recognition with artificial neural network backpropagation concept. The artificial neural network use sigmoid activation function to all layer. Steps to the extraction, fi...
详细信息
This paper presents a comparison among different strategies to coordinate the use of heterogeneous wireless sensors aimed for area surveillance. The heterogeneity among the sensor nodes is related to their sensing and...
详细信息
News event modeling and tracking in the social web is the task of discovering which news events individuals in social communities are most interested in, how much discussion these events generate and tracking these di...
详细信息
News event modeling and tracking in the social web is the task of discovering which news events individuals in social communities are most interested in, how much discussion these events generate and tracking these discussions over time. The task could provide informative summaries on what has happened in the real world, yield important knowledge on what are the most important events from the crowd's perspective and reveal their temporal evolutionary trends. Latent Dirichlet Allocation (LDA) has been used intensively for modeling and tracking events (or topics) in text streams. However, the event models discovered by this bottom-up approach have limitations such as a lack of semantic correspondence to real world events. Besides, they do not scale well to large datasets. This paper proposes a novel latent Dirichlet framework for event modeling and tracking. Our approach takes into account ontological knowledge on events that exist in the real world to guide the modeling and tracking processes. Therefore, event models extracted from the social web by our approach are always meaningful and semantically match with real world events. Practically, our approach requires only a single scan over the dataset to model and track events and hence scales well with dataset size.
Computational energy consumption of the processing elements (PEs) of a NoC can be significantly reduced by scaling down their voltage levels. This creates clusters of adjacent PEs operating at the same voltage level, ...
详细信息
Power consumption can be significantly reduced in Systems-on-Chip (SoC) by scaling down the voltage levels of the Processing Elements (PEs). The power efficiency of this Voltage Islanding technique comes at the cost o...
详细信息
ISBN:
(纸本)9783981080162
Power consumption can be significantly reduced in Systems-on-Chip (SoC) by scaling down the voltage levels of the Processing Elements (PEs). The power efficiency of this Voltage Islanding technique comes at the cost of energy and area overhead due to the level shifters between voltage islands. Moreover, from the physical design perspective it is not desirable to have an excessive number of voltage islands on the chip. Considering voltage islanding at an early phase of design as during floorplanning of the PEs can address various of these issues. In this paper, we propose a new cost function for the floorplanning objective different from the traditional floorplanning objective. The new cost function not only includes the overall area requirement, but also incorporates the overall power consumption and the design constraint imposed on the maximum number of voltage islands. We propose a greedy heuristic based on the proposed cost function for the floorplanning of the PEs with several voltage islands. Experimental results using benchmark data study the effect of several parameters on the outcome of the heuristic. It is evident from the results that power consumption can be significantly reduced using our algorithm without significant area overhead. The area obtained from the heuristic is also compared with the optimal, and found to be within 4% of the optimal on average, when area minimization is given the priority.
An important problem in Wireless Sensor Networks (WSN) is the occurrence of failures that lead to the disconnection of parts of the network, compromising the final results achieved by the WSN operation. A way to overc...
详细信息
Errors in diagnosing the disease is a critical risk that must be faced by any person giving treatment to the hospital. Medical treatment can not always be done with perfect accuracy. Lung cancer is one of the most dea...
详细信息
Errors in diagnosing the disease is a critical risk that must be faced by any person giving treatment to the hospital. Medical treatment can not always be done with perfect accuracy. Lung cancer is one of the most deadly disease that prone to misdiagnose. In general, some practitioners tend to “read” cancer in x-ray rontgen image as tumor this could be fatal. To generate a diagnose, a general practitioner use three kind of examination i.e : patient History, Radio logic examination, phisical examination. In this paper, Gray color for image indexing and retrieval are investigated. The features are derived based on the statistical distribution of Harralick feature from image sample. By utilizing the proposed invariant features, the similarity measure between query and database images provides reliable retrieval results.
A well know problem in the Wireless Sensor Network (WSN) research area is the usage of appropriate strategies to setup the sensor nodes such that they may accomplish sensing missions. This problem refers to the select...
详细信息
Artificial Immune Systems (AISs) are composed of techniques inspired by immunology. The clonal selection principle ensures the organism adaptation to fight invading antigens by an immune response activated by the bind...
详细信息
Artificial Immune Systems (AISs) are composed of techniques inspired by immunology. The clonal selection principle ensures the organism adaptation to fight invading antigens by an immune response activated by the binding of antigens and antibodies. As an immune response can be elicited even when the binding between an antigen and an antibody is not perfect, an approximate binding might suffice, and a Fuzzy Logic mechanism might be the most appropriate mechanism to control such process. This paper presents a novel hybrid model based on concepts of Immune and Fuzzy Systems with applications to pattern recognition problems. The preliminary results obtained here suggest the proposed model is a promising pattern recognition tool.
暂无评论