Yohan J. Roh

Hierarchically organized skew-tolerant histograms for geographic data objects

Histograms have been widely used for fast estimation of query result sizes in query optimization. In this paper, we propose a new histogram method, called the Skew-Tolerant Histogram (STHistogram) for two or three dimensional geographic data objects that are used in many real-world applications in practice. The proposed method provides a significantly enhanced accuracy in a robust manner even for the data set that has a highly skewed distribution. Our method detects hotspots present in various parts of a data set and exploits them in organizing histogram buckets. For this purpose, we first define the concept of a hotspot, and provide an algorithm that efficiently extracts hotspots from the given data set. Then, we present our histogram construction method that utilizes hotspot information. We also describe how to estimate query result sizes by using the proposed histogram. We show through extensive performance experiments that the proposed method provides better performance than other existing methods.

Rule-based Dynamic Business Process Modification and Adaptation

The advances in ubiquitous computing technologies have given rise to the u-Business, where we obtain business-related context information from ubiquitous devices to adapt to the changes of business environments in a timely manner. In these rapidly changing u-Business environments, we need a mechanism to modify business process schemas, which represent a group of business activities arranged in a meaningful way to give values to the u-Business. We also need a way to migrate running process instances to their new business process schemas. In this paper, we propose a rule-driven approach for dynamic business process schema modification and instance adaptation. The proposed approach uses rules to describe schema modification operations to be applied and rule execution conditions. Business environment changes are encoded in these rules, thus timely schema modification is possible. We also propose a system architecture for realizing this rule-driven approach. We have designed the system architecture by utilizing existing technologies so that the system can be easily implemented.

Semantic Business Registry Information Model

For the intelligent business process management, ebXML registry needs to efficiently store and process semantic information. Even though ebXML registry provides two basic information models to define and associate semantics between registry objects; e.g., classification information model and association information model, they cannot express all semantics of the standard Web ontology language (OWL) due to their generic differences in design goals and usage patterns. To support OWL semantics in ebXML registries is to provide a foundation for intelligent information processing in ebXML registries. In this paper, we address how to embed OWL semantics in the ebXML registry. To express OWL constructs in ebXML registries, we propose a new ebXML registry information model, i.e., a semantic information model (SIM).

Energy-efficient two-dimensional skyline query processing in wireless sensor networks

Wireless sensor networks have been widely used for environment monitoring and event surveillance. In recent years, the skyline query, as an important operator for multiple-criteria decision making, has received much attention because of its many real-world applications. In this paper, we propose a new filter-based method, called SKYFILTER, for skyline query processing. Our method provides an enhanced efficiency by reduction of the total wireless communication between sensor nodes. In addition, the proposed method also provides a better performance than other existing methods with respect to the computational efficiency of skyline computation in sensor nodes. In this paper, we first present a new multiple filter-based technique to reduce wireless communication. Then, we provide a new merge sort-based algorithm for efficient skyline computation in sensor nodes. We show through performance experiments and algorithm analysis that the proposed method generally provides a better performance than other existing methods in terms of the wireless communication cost and computational efficiency.

4.1 A 640×480 dynamic vision sensor with a 9µm pixel and 300Meps address-event representation

We report a VGA dynamic vision sensor (DVS) with a 9µm pixel, developed through a digital as well as an analog implementation. DVS systems in the literature try to increase spatial resolution up to QVGA [1–2] and data rates up to 50 million events per second (Meps) (self-acknowledged) [3], but they are still inadequate for high-performance applications such as gesture recognition, drones, automotive, etc. Moreover, the smallest reported pixel of 18.5µm is too large for economical mass production [3]. This paper reports a 640×480 VGA-resolution DVS system with a 9µm pixel pitch supporting a data rate of 300Meps for sufficient event transfer in spite of higher resolution. Maintaining acceptable pixel performance, the pixel circuitry is carefully designed and optimized using a BSI CIS process. To acquire data (i.e., pixel events) at high speed even with high resolution (e.g., VGA), a fully synthesized word-serial group address-event representation (G-AER) is implemented, which handles massive events in parallel by binding neighboring 8 pixels into a group. In addition, a 10b programmable bias generator dedicated to a DVS system provides easy controllability of pixel biases and event thresholds.

STHist-C: a highly accurate cluster-based histogram for two and three dimensional geographic data points

Histograms have been widely used for estimating selectivity in query optimization. In this paper, we propose a new histogram construction method for geographic data objects that are used in many real-world applications. The proposed method is based on analyses and utilization of clusters of objects that exist in a given data set, to build histograms with significantly enhanced accuracy. Our philosophy in allocating the histogram buckets is to allocate them to the subspaces that properly capture object clusters. Therefore, we first propose a procedure to find the centers of object clusters. Then, we propose an algorithm to construct the histogram buckets from these centers. The buckets are initialized from the clusters’ centers, then expanded to cover the clusters. Best expansion plans are chosen based on a notion of skewness gain. Results from extensive experiments using real-life data sets demonstrate that the proposed method can really improve the accuracy of the histograms further, when compared with the current state of the art histogram construction method for geographic data objects.

Efficient construction of histograms for multidimensional data using quad-trees

Histograms can be useful in estimating the selectivity of queries in areas such as database query optimization and data exploration. In this paper, we propose a new histogram method for multidimensional data, called the Q-Histogram, based on the use of the quad-tree, which is a popular index structure for multidimensional data sets. The use of the compact representation of the target data obtainable from the quad-tree allows a fast construction of a histogram with the minimum number of scanning, i.e., only one scanning, of the underlying data. In addition to the advantage of computation time, the proposed method also provides a better performance than other existing methods with respect to the quality of selectivity estimation. We present a new measure of data skew for a histogram bucket, called the weighted bucket skew. Then, we provide an effective technique for skew-tolerant organization of histograms. Finally, we compare the accuracy and efficiency of the proposed method with other existing methods using both real-life data sets and synthetic data sets. The results of experiments show that the proposed method generally provides a better performance than other existing methods in terms of accuracy as well as computational efficiency.

Content-based multipath routing for sensor networks

In wireless sensor networks, in-network processing of aggregation queries has been an important technique to reduce energy consumption in wireless communication, which is a main source of energy consumption in sensor devices. In-network processing is typically guided by an aggregation tree, where each node forwards partially computed aggregates to its parent. In this paper we consider a routing method for grouped aggregation queries, where sensor readings are divided into disjoint groups according to their values and aggregates are computed for each group. For this type of queries, multipath routing, where each node forwards different aggregates to different nodes, can lead to more efficient in-network processing. However, no multipath routing protocol for efficient in-network processing of grouped aggregation queries has been proposed thus far. In this paper we propose a new routing protocol, called Content-based Multipath Routing (CMR), for efficient in-network processing of grouped aggregation queries. CMR employs multipath routing, and each node forwards partially computed aggregates along different paths based on the contents of the aggregates. The experimental results show that CMR outperforms the existing aggregation tree-based routing protocols.

Performance improvement of deep learning based gesture recognition using spatiotemporal demosaicing technique

We propose a novel method for the demosaicing of event-based images that offers substantial performance improvement of far-distance gesture recognition based on deep Convolutional Neural Network. Unlike the conventional demosaicing technique using the spatial color interpolation of Bayer patterns, our new approach utilizes spatiotemporal correlation between pixel arrays, whereby timestamps of high-resolution pixels are efficiently generated in real-time from the event data. In this paper, we describe this new method and evaluate its performance with a hand motion recognition task.