In the context of support vector machines, feature selection is motivated mainly by the consideration of classification speed and generalization ability. Sensitivity analysis of MLP and RBF has already been successful...
详细信息
ISBN:
(纸本)0780384032
In the context of support vector machines, feature selection is motivated mainly by the consideration of classification speed and generalization ability. Sensitivity analysis of MLP and RBF has already been successfully applied in feature subset selection. In this paper, we present a novel feature selection method for Support Vector Machines (SVMs) using the sensitivity analysis of SVMs, which is defined as the deviation of separation margin with respect to the perturbation of given feature. The method we proposed can directly be applied to multi-class SVMs. Our experiments validate that the proposed strategy produces satisfactory results both on artificial and real-world data.
The Low Frequency-Asynchronous Switch Design (the LF-ASD) was introduced as a direct asynchronous Brain Computer Interface (BCI) system. An asynchronous BCI system is activated only when a user intends control and mai...
详细信息
The curvature scale space (CSS) technique is one of the key techniques of the MPEG-7 international standard in Computer Vision and Image Processing It was selected as a contour shape descriptor for MPEG-7 after substa...
详细信息
ISBN:
(纸本)0769520782
The curvature scale space (CSS) technique is one of the key techniques of the MPEG-7 international standard in Computer Vision and Image Processing It was selected as a contour shape descriptor for MPEG-7 after substantial and comprehensive testing However to compute a CSS image in general needs to wait a long time. This is very disadvantageous when the CSS technique is applied to an object recognition system to perform real-time recognition. In order to solve this bottleneck problem, a hybrid method for fast computing the CSS image is proposed in the present paper In the method, firstly the curve is evolved in low scale space, and after image noise is suppressed then the curvature is evolved directly Numerical experiments show that the hybrid method can perform equally well as the existing method. It is suitable for recognizing a noisy curve of arbitrary shape at any scale or orientation. On the other hand, the hybrid method only requires 1/3 similar to 1/5 CPU time of the existing one. As a result, the CSS technique is improved significantly for real-time recognition.
In this article three methods are presented to perform the center of gravity defuzzification method: one well-known method, the discretisation method, and two new methods, the slope-based method and the modified trans...
详细信息
ISBN:
(纸本)0780383532
In this article three methods are presented to perform the center of gravity defuzzification method: one well-known method, the discretisation method, and two new methods, the slope-based method and the modified transformation function method. The methods are worked out for trapezoidal membership functions forming a fuzzy (Ruspini) partition. Experimental results show that the newly introduced methods exhibit excellent accuracy at an extremely low computational cost compared to the widely applied discretisation method.
Fielded mobile robot systems will inevitably suffer hardware and software failures. Failures in a single subsystem can often disable the entire robot, especially if the controlling application does not consider such f...
详细信息
ISBN:
(纸本)0819455628
Fielded mobile robot systems will inevitably suffer hardware and software failures. Failures in a single subsystem can often disable the entire robot, especially if the controlling application does not consider such failures. Often simple measures, such as a software restart or the use of a secondary sensor, can solve the problem. However, these fixes must generally be applied by a human expert, who might not be present in the field. In this paper, we describe a recovery-oriented framework for mobile robot applications which addresses this problem in two ways. First, fault isolation automatically provides graceful degradation of the overall system as individual software and hardware components fail. In addition, subsystems are monitored for known failure modes or aberrant behavior. The framework responds to detected or immanent failures by restarting or replacing the suspect component in a manner transparent to the application programmer and the robot's operator.
This paper deals with the development of an inspection method for the evaluation of synchronous groupware in e-Learning. Distributed groupware are software applications that are able to facilitate collaboration among ...
详细信息
A locally decodable code (LDC) encodes n-bit strings x in m-bit codewords C(x) in such a way that one can recover any bit xi from a corrupted codeword by querying only a few bits of that word. We use a quantum argumen...
详细信息
A locally decodable code (LDC) encodes n-bit strings x in m-bit codewords C(x) in such a way that one can recover any bit xi from a corrupted codeword by querying only a few bits of that word. We use a quantum argument to prove that LDCs with 2 classical queries require exponential length: m = 2(Omega(17)). Previously, this was known only for linear codes (Goldreich et al., in: proceedings of 17th IEEE Conference on Computation Complexity, 2002, pp. 175-183). The proof proceeds by showing that a 2-query LDC can be decoded with a single quantum query, when defined in an appropriate sense. It goes on to establish an exponential lower bound on any '1-query locally quantum-decodable code'. We extend our lower bounds to non-binary alphabets and also somewhat improve the polynomial lower bounds by Katz and Trevisan for LDCs with more than 2 queries. Furthermore, we show that q quantum queries allow more succinct LDCs than the best known LDCs with q classical queries. Finally, we give new classical lower bounds and quantum upper bounds for the setting of private information retrieval. In particular, we exhibit a quantum 2-server private information retrieval (PIR) scheme with O(n(3/10)) qubits of communication, beating the O(n(1/3)) bits of communication of the best known classical 2-server PIR. (C) 2004 Elsevier Inc. All rights reserved.
State-of-the-art Massively Multiplayer Games such as EverQuest and Ultima Online are currently implemented as client-server systems. Although this approach allows the development of commercially viable MMG services, t...
详细信息
ISBN:
(纸本)0769522327
State-of-the-art Massively Multiplayer Games such as EverQuest and Ultima Online are currently implemented as client-server systems. Although this approach allows the development of commercially viable MMG services, the costs associated with running a scalable client-server MMG service are often too high for small companies or research projects. This paper proposes FreeMMG, a mixed peer-to-peer and client-server approach to the distribution aspect of MMGs. It is argued that the FreeMMG model supports scalable, cheat-resistant, massively multiplayer real-time strategy (RTS) games using a lightweight server that delegates the bulk of the game simulation to the clients. A working prototype game called FreeMMG Wizards is presented, together with some preliminary scalability test results featuring up to 300 simulated game clients connected to a FreeMMG server The results show that the measured server traffic can be considered very low if compared with more centralized alternatives.
Distributed Interactive Applications (DIAs) have been gaining commercial success in recent years due to the widespread appeal of networked multiplayer computer games. Within these games, participants interact with eac...
详细信息
ISBN:
(纸本)0769522327
Distributed Interactive Applications (DIAs) have been gaining commercial success in recent years due to the widespread appeal of networked multiplayer computer games. Within these games, participants interact with each other and their environment, producing complex behavioural patterns that evolve over time. These patterns are non-linear, and often appear to exhibit dependencies under certain conditions. In this paper, we analyse the behavioural patterns of two users interacting in a DIA. Our motivation behind this analysis is to construct models of user behaviour that can be used within Entity-State-Update (ESU) mechanisms. By representing their behaviour as time-series datasets, we investigate the use of simple statistical dependence measures to help partition the datasets and identify three different types of behavioural states exhibited by the two users. It is our intention that future research on ESU mechanisms can utilize this behavioural partitioning to reduce the network traffic in a DIA based on a hybrid-model approach.
In mobile Internet applications, data can be transcoded, updated, and transferred across heterogenous clients. The problem then arises where updates made in the context of an initial transcoding results in content too...
详细信息
ISBN:
(纸本)1581137931
In mobile Internet applications, data can be transcoded, updated, and transferred across heterogenous clients. The problem then arises where updates made in the context of an initial transcoding results in content too stringently transcoded for subsequent clients, thereby causing loss of semantic value. We solve this problem by suggesting that the updates themselves can be transformed so that they can be applied directly to the original data instead of to the transcoded data;this approach allows the data to preserve as much semantic value as possible across all heterogeneous clients without unnecessary transcoding artifacts. We define reconciliation rules that can govern the interaction between client updates and transcoding, demonstrate a complete middleware architecture that supports our methodology, and provide two case studies using content-transferring applications. We show that our resulting middleware system executes our reconciliation approach with acceptable latency (under 5 seconds for 200 kbytes of layered content), good scalability, and well-organised modularity.
暂无评论