The classification task usually works with flat and batch learners, assuming problems as stationary and without relations between class labels. Nevertheless, several real-world problems do not assume these premises, i.e., data have labels organized hierarchically and are made available in streaming fashion, meaning that their behavior can drift over time. Existing studies on hierarchical classification do not consider data streams as input of their process, and thus, data is assumed as stationary and handled through batch learners. The same can be said about works on streaming data, as the hierarchical classification is overlooked. Studies concerning each area individually are promising, yet, do not tackle their intersection. This study analyzes the main characteristics of the state-of-the-art works on hierarchical classification for streaming data concerning five aspects: (i) problems tackled, (ii) datasets, (iii) algorithms, (iv) evaluation metrics, and (v) research gaps in the area. We performed a systematic literature review of primary studies and retrieved 3,722 papers, of which 42 were identified as relevant and used to answer the aforementioned research questions. We found that the problems handled by hierarchical classification of data streams include mainly classification of images, human activities, texts, and audio; the datasets are mostly created or synthetic data; the algorithms and evaluation metrics are well-known techniques or based on those; and research gaps are related to dynamic context, data complexity, and computational resources constraints. We also provide implications for future research and experiments to consider common characteristics shared amongst hierarchical classification and data stream classification.
This paper shows that breaking the barrier of 1 triangle/clock rasterization rate for microtriangles in modern GPU architectures in an efficient way is possible. The fixed throughput of the special purpose culling and triangle setup stages of the classic pipeline limits the GPU scalability to rasterize many triangles in parallel when these cover very few pixels. In contrast, the shader core counts and increasing GFLOPs in modern GPUs clearly suggests parallelizing this computation entirely across multiple shader threads, making use of the powerful wide-ALU instructions. In this paper, we present a very efficient SIMD-like rasterization code targeted at very small triangles that scales very well with the number of shader cores and has higher performance than traditional edge equation based algorithms. We have extended the ATTILA GPU shader ISA (del Barrioet al. in IEEE International Symposium on Performance Analysis of Systems and Software, pp. 231–241, 2006) with two fixed point instructions to meet the rasterization precision requirement. This paper also introduces a novel subpixel Bounding Box size optimization that adjusts the bounds much more finely, which is critical for small triangles, and doubles the 2×2-pixel stamp test efficiency. The proposed shader rasterization program can run on top of the original pixel shader program in such a way that selected fragments are rasterized, attribute interpolated and pixel shaded in the same pass. Our results show that our technique yields better performance than a classic rasterizer at 8 or more shader cores, with speedups as high as 4× for 16 shader cores. 相似文献
Most simulation models of electric machines use the coupled circuit approach, where the machine is considered as an electric circuit element with time-varying inductances (abc model) or with constant inductances (dq0 model). On the other hand, the rotating magnetic field approach, which considers the electric machine as two groups of windings producing rotating magnetic fields and can give insight into internal phenomena of the machines, has not yet received much attention in electric machines modeling, especially for machine transient analysis. Based on the rotating magnetic field approach, this paper presents a transient model of the induction machine including main flux saturation effect. Based on the direct computation of the magnetizing fluxes of all machine windings, the model represents instantaneous main flux saturation by simply introducing a main flux saturation factor. No iteration process is involved to incorporate the saturation effects. The model combines the advantages of the dq0 and abc models advantages, such as rapid computation time and nonsymmetrical conditions simulation, respectively. The simulation results and the experimental tests show advantages and verification of the model. 相似文献
Oxygen sensing films were synthesized by a chemical conjugation of functional platinum porphyrin probes in silica gel, polystyrene (PS), and poly(2-hydroxyethyl methacrylate) (PHEMA) matrices. Responses of the sensing films to gaseous oxygen and dissolved oxygen were studied and the influence of the matrices on the sensing behaviors was investigated. Silica gel films had the highest fluorescence intensity ratio from deoxygenated to oxygenated environments and the fastest response time to oxygen. PHEMA films had no response to gaseous oxygen, but had greater sensitivity and a faster response time for dissolved oxygen than those of PS films. The influence of matrices on oxygen response, sensitivity and response time was discussed. The influence is most likely attributed to the oxygen diffusion abilities of the matrices. Since the probes were chemically immobilized in the matrices, no leaching of the probes was observed from the sensing films when applied in aqueous environment. One sensing film made from the PHEMA matrix was used to preliminarily monitor the oxygen consumption of Escherichia coli (E. coli) bacteria. E. coli cell density and antibiotics ampicillin concentration dependent oxygen consumption was observed, indicating the potential application of the oxygen sensing film for biological application. 相似文献
Web applications are fast becoming more widespread, larger, more interactive, and more essential to the international use of computers. It is well understood that web applications must be highly dependable, and as a field we are just now beginning to understand how to model and test Web applications. One straightforward technique is to model Web applications as finite state machines. However, large numbers of input fields, input choices and the ability to enter values in any order combine to create a state space explosion problem. This paper evaluates a solution that uses constraints on the inputs to reduce the number of transitions, thus compressing the FSM. The paper presents an analysis of the potential savings of the compression technique and reports actual savings from two case studies. 相似文献
A systematic approach to parameter-dependent control synthesis of a high-speed supercavitation vehicle (HSSV) is presented. The aim of the control design is to provide robust reference tracking across a large flight envelope, while directly accounting for the interaction of liquid and gas phases with the vehicle. A nonlinear dynamic HSSV model is presented and discussed relative to the actual vehicle. A linear, parameter-varying (LPV) controller is synthesized for angle rate tracking in the presence of model uncertainty. The control design takes advantage of coupling in the governing equations to achieve improved performance. Multiple LPV controllers synthesized for smaller overlapping regions of the parameter space are blended together, providing a single controller for the full flight envelope. Time-domain simulations implemented on high-fidelity simulations, provide insight into the performance and robustness of the proposed scheme. 相似文献
Squeeze-film damping on microresonators is a significant damping source even when the surrounding gas is highly rarefied. This article presents a general modeling approach based on Monte Carlo (MC) simulations for the prediction of squeeze-film damping on resonators in the free-molecule regime. The generality of the approach is demonstrated in its capability of simulating resonators of any shape and with any accommodation coefficient. The approach is validated using both the analytical results of the free-space damping and the experimental data of the squeeze-film damping on a clamped–clamped plate resonator oscillating at its first flexure mode. The effect of oscillation modes on the quality factor of the resonator has also been studied and semi-analytical approximate models for the squeeze-film damping with diffuse collisions have been developed. 相似文献
Intense competition in the UK charity fundraising market and the near continuous need for charities to acquire new supporters, reactivate lapsed donors and encourage first time donors to make further gifts has resulted in charities needing repeatedly to initiate novel fundraising campaigns. The starting point for an innovative campaign is an idea or set of ideas that will attract public attention, differentiate a charity from other fundraising organizations and distinguish the charity's current marketing activities from those undertaken in the past. An important but neglected issue in the charity idea creation field is the question of where and how fundraising organizations source their ideas for new campaigns. This empirical study identified (i) the main sources of fresh ideas for fundraising employed by a sample of 417 British charities, and (ii) certain antecedents of a charity's use of various categories of sources. Potential connections between particular types of idea source and managerial satisfaction with both the creative quality of the ideas obtained and the financial revenues derived from these ideas were also examined. 相似文献
Interior gateway routing protocols like Open Shortest Path First (OSPF) and Distributed Exponentially Weighted Flow Splitting (DEFT) send flow through forward links toward the destination node. OSPF routes only on shortest‐weight paths, whereas DEFT sends flow on all forward links, but with an exponential penalty on longer paths. Finding suitable weights for these protocols is known as the weight setting problem (WSP). In this paper, we present a biased random‐key genetic algorithm for WSP using both protocols. The algorithm uses dynamic flow and dynamic shortest path computations. We report computational experiments that show that DEFT achieves less network congestion when compared with OSPF, while, however, yielding larger delays. 相似文献
Given a distributed system of \(n\) balls and \(n\) bins, how evenly can we distribute the balls to the bins, minimizing communication? The fastest non-adaptive and symmetric algorithm achieving a constant maximum bin load requires \(\varTheta (\log \log n)\) rounds, and any such algorithm running for \(r\in {\mathcal {O}}(1)\) rounds incurs a bin load of \(\varOmega ((\log n/\log \log n)^{1/r})\). In this work, we explore the fundamental limits of the general problem. We present a simple adaptive symmetric algorithm that achieves a bin load of 2 in \(\log ^* n+{\mathcal {O}}(1)\) communication rounds using \({\mathcal {O}}(n)\) messages in total. Our main result, however, is a matching lower bound of \((1-o(1))\log ^* n\) on the time complexity of symmetric algorithms that guarantee small bin loads. The essential preconditions of the proof are (i) a limit of \({\mathcal {O}}(n)\) on the total number of messages sent by the algorithm and (ii) anonymity of bins, i.e., the port numberings of balls need not be globally consistent. In order to show that our technique yields indeed tight bounds, we provide for each assumption an algorithm violating it, in turn achieving a constant maximum bin load in constant time. 相似文献