Join is one of the most important operations in data analytics systems. Prior works focus mainly on join optimization using GPUs, but little is known about performance impact on the MICs. In order to investigate poten...
详细信息
This paper proposes a novel method for cross-modal retrieval. Different from vector (text)-to-vector (image) framework of the traditional cross-modal methods, we adopt a vector (text)-to-matrix (image) framework. We a...
详细信息
ISBN:
(纸本)9781509060689
This paper proposes a novel method for cross-modal retrieval. Different from vector (text)-to-vector (image) framework of the traditional cross-modal methods, we adopt a vector (text)-to-matrix (image) framework. We assume that compared with vectors, matrices can directly represent images and characterize the structure of feature space. Furthermore, we propose a Metric based on Multi-order spaces (MMs). Multi-order statistic features are used to represent images for enriching the semantic information, and metrics among the multi-spaces are jointly learned to measure the similarity between two different modalities. Specifically, there are three steps for MMs. First, we jointly use the bags of visual features (zero-order), mean (first-order) and covariance (second-order) to characterize each image. Second, considering that covariance matrices and vectors lie on a Riemannian manifold and an Euclidean space respectively, we embed multi-order spaces into their corresponding Hilbert spaces to reduce the heterogeneity among the original spaces. Finally, the similarity between two different modalities can be measured by learning multiple transformations from the different Hilbert spaces to a common subspace. The performance of the proposed method over the state-of-the-art has been demonstrated through the experiments on two public datasets.
Till now, a large variety of researchers have carried out lots of efforts on object-oriented and UML model metrics from different views. They put forward numerous of metrics and carried out some series of theoretical ...
详细信息
Till now, a large variety of researchers have carried out lots of efforts on object-oriented and UML model metrics from different views. They put forward numerous of metrics and carried out some series of theoretical and experimental verifications on understandability, analyzability, maintainability, fault-proneness, change-proneness and reuse. However, there is no formal semantic specification for UML model metrics, which may lead to potential semantic inconsistency and ambiguity. To solve this problem, this paper provided formalization for UML model metrics at the level of UML Meta models. This formalization can not only help people to understand the meaning of UML model metrics, but also can be used in the application domain of UML model metrics in a more rigorous way.
Multiple sensors and sensor fusion are commonly used to get more accurate information. The intuitive method to store multi-sensory data is using uncertain database because the sensors are not precise enough. Hence, li...
详细信息
Multiple sensors and sensor fusion are commonly used to get more accurate information. The intuitive method to store multi-sensory data is using uncertain database because the sensors are not precise enough. Hence, like the top-k queries in traditional database, the top-k queries in uncertain databases are quite popular and useful due to its wide application. Although there are lots of top-k query semantics, most of them return tuples, which does not make sense in some cases. We define two novel kinds of top-k query semantics in uncertain database, Uncertain x-kRanks queries (U-x-kRanks) and Global x-Top-k queries (G-x-Top-k), which return k x-tuples according to the score and the confidence of alternatives in x-tuples, respectively. Moreover, in order to reduce the search space, we present an efficient algorithm to process U-x-kRanks queries and G-x-Top-k queries. Comprehensive experiments and analysis on different artificial data sets demonstrate the effectiveness of the proposed strategies.
Online customer review is considered as a significant informative resource which is useful for both potential customer and product manufacturers. As a result, it is one of the most challenging tasks to mine customer r...
详细信息
Online customer review is considered as a significant informative resource which is useful for both potential customer and product manufacturers. As a result, it is one of the most challenging tasks to mine customer reviews automatically and to provide users with opinion summary. Product features and opinion word play the most important roles in the customers' opinions mining. In this paper, we dedicate our work to opinion word mining. We proposed an approach for opinion word identification based on the association rule mining algorithm. The method makes full use of co-occurrence syntactic characteristic between product features and opinion word. Firstly, the product feature is identified by two-stage filtering scheme, and secondly the opinion word is extracted through association rule mining. The final experiment results show that the proposed method could not only obtain the product features related to domain characteristics, but identify the opinion word effectively. Meanwhile, our approach possesses much higher precision and recall than Hu's work.
With the aim of resolving the issue of cluster analysis more precisely and validly, a new approach was proposed based on biogeography-based optimization (abbreviated as BBO) algorithm. (Method) First, we reformulated ...
详细信息
This paper proposes an effective fusion of Analytic Hierarchy Process (AHP) and Grey Relational Analysis (GRA) approach for the risk evaluation in Mobile Commerce (MC) development. The hybrid method employs the comple...
详细信息
Outlier detection is a crucial part of robust evaluation for crowd-sourceable assessment of Quality of Experience (QoE) and has attracted much attention in recent years. In this paper, we propose some simple and fast ...
详细信息
This paper summarizes our efforts for the first time participation in the Violent Scene Detection subtask of the MediaEval 2015 Affective Impact of Movies Task. We build violent scene detectors using both audio and vi...
This paper summarizes our efforts for the first time participation in the Violent Scene Detection subtask of the MediaEval 2015 Affective Impact of Movies Task. We build violent scene detectors using both audio and visual cues. In particular, the audio cue is represented by bag-of-audio-words with fisher vector encoding. The visual cue is exploited by extracting CNN features from video frames. The detectors are implemented using two-class linear SVM classifiers. Evaluation shows that the audio detectors and the visual detectors are comparable and complementary to each other. Among our submissions, multi-modal late fusion leads to the best performance.
This abstract paper sketches our research towards Struc-tured Semantic Embedding of multimedia data. Though a tag may have multiple senses with completely different visual imagery, current semantic embedding methods r...
详细信息
暂无评论