Search

Indexes are crucial for the efficient processing of database workloads and an appropriately selected set of indexes can drastically improve query processing performance. However, the selection of beneficial indexes is a non-trivial problem and still challenging. Recent work in deep reinforcement learning (DRL) may bring a new perspective on this problem. In this paper, we studied the index selection problem in the context of reinforcement learning and proposed an end-to-end DRL-based index selection framework. The framework poses the index selection problem as a series of 1-step single index recommendation tasks and can learn from data. Unlike most existing DRL-based index selection solutions that focus on selecting single-column indexes, our framework can recommend both single-column and multi-column indexes for the database. A set of comparative experiments with existing solutions was conducted to demonstrate the effectiveness of our proposed method.

With the development of intelligent manufacturing industry, the management of tool data in machine tool processing is becoming more and more important. Due to the richness of machine tool data and the complexity of relationships in them, it’s hard for a traditional relational database to manage the tool data. Therefore, a new method should be proposed to manage these data in a better way. In this work, we propose a tool data modeling method based on the object deputy model, which utilizes the characteristics of the class and the objects to express the meaning of the tool data and the various semantic constraints on them. Unlike the traditional relational model, objects are connected with a two-way pointer in the object deputy model where an object can have one or more deputy objects that inherit the properties and methods of the source object, and the deputy objects can have their own properties and methods. Besides, the two-way pointer between the source class and its deputy class makes the cross-class query easier in two aspects: One is to make complex queries expressed in intuitive statements, and the other is to improve query efficiency. We implemented and evaluated our model on an object deputy database. Experiments show that our method is better than the traditional relational ones.

Spatial data has been widely used in many areas. Meanwhile, the increasingly simplified data access and manipulation methods in spatial data related applications make it important to verify data truthfulness. In this paper, we propose a reversible fragile watermarking scheme for the content authentication of spatial data. The proposed scheme embeds (detects) four kinds of watermarks, i.e., group, object, vertex-local and vertex-global watermarks, into (from) each object (vertex) by altering the digits of the coordinates after the LSD (Least Significant Digit). According to the detection results of those four kinds of watermarks, the scheme can not only locate tampers, but also recognize modification types on both object and vertex levels. In addition, the proposed scheme is reversible, that is the original data can be restored from the watermarked data. The feature of reversibility makes the scheme suitable for applications which require high-precision data.

With the development of the positioning technology, studies on trajectories have been growing rapidly in the past decades. As a fundamental part involved in trajectory recommendation and prediction, trajectory similarity has attracted considerable attention from researchers. However, most existing works focus on raw trajectory similarity by comparing their shapes, while very few works study semantic trajectory similarity and none of them take all the geographical, semantic, and timestamp information into consideration. In this paper, we model semantic trajectories based on road networks considering all these information, and propose a Constrained Time-based Common Parts (CTCP) approach to measure the similarity. Since the strict time constraint in CTCP may lead to many zero values, we further propose an improved Weighted Constrained Time-based Common Parts (WCTCP) method by relaxing the time constraint to measure the similarity more accurately. We conducted extensive performance studies on real datasets to confirm the effectiveness of our approaches.

As the most important geographical data, 2D vector maps have been widely used in various areas. Meanwhile, the increasingly simplified data access and manipulation methods in geographical data related applications make it important to verify data truthfulness. In this paper, we propose a fragile watermarking scheme for the content authentication of 2D vector maps. The scheme embeds (detects) two kinds of watermarks, i.e., group watermark and object watermark, into (from) each object. Based on the patterns extracted from the detection results, the scheme can not only detect and locate data modifications, but also characterize modification types. Both theoretical analysis and experimental evaluation are carried out to verify the fragility of the proposed scheme.

More and more trajectory data are available as streams due to the unprecedented prevalence of mobile positioning devices. Meanwhile, an increasing number of applications are designed to be dependent on real-time trajectory streams. Therefore, the protection of ownership rights over such data becomes a necessity. In this paper, we propose an online watermarking scheme that can be used for the rights protection of trajectory streams. The scheme works in a finite window, single-pass streaming model. It embeds watermark by modifying feature distances extracted from the streams. The fact that these feature distances can be recovered ensures a consistent overlap between the recovered watermark and the embedded one. Experimental results verify the robustness of the scheme against domain-specific attacks, including geometric transformations, noise addition, trajectory segmentation and compression.