Big data mining is well known to be an important task for data science, because it can provide useful observations and new knowledge hidden in given large datasets. Proximity-based data analysis is particularly utilized in many real-life applications. In such analysis, the distances to k nearest neighbors are usually employed, thus its main bottleneck is derived from data retrieval. Much efforts have been made to improve the efficiency of these analyses. However, they still incur large costs, because they essentially need many data accesses. To avoid this issue, we propose a machine-learning technique that quickly and accurately estimates the k-NN distances (i.e., distances to the k nearest neighbors) of a given query. We train a fully connected neural network model and utilize pivots to achieve accurate estimation. Our model is designed to have useful advantages: it infers distances to the k-NNs at a time, its inference time is O(1) (no data accesses are incurred), but it keeps high accuracy. Our experimental results and case studies on real datasets demonstrate the efficiency and effectiveness of our solution.

Due to the proliferation of GPS-enabled mobile devices and IoT environments, location-based services are generating a large number of objects that contain both spatial and keyword information, and spatial-keyword databases are receiving much attention. This paper addresses the problem of top-k spatial-keyword similarity join, which outputs k object pairs with the highest similarity. This query is a primitive operator for important applications, including duplicate detection, recommendation, and clustering.

This work proposes a method for estimating dynamics on graph by using dynamic mode decomposition (DMD) and sparse approximation with graph filter banks (GFBs). The motivation of introducing DMD on graph is to predict multi-point river water levels for forecasting river flood and giving proper evacuation warnings. The proposed method represents a spatio-temporal variation of physical quantities on a graph as a time-evolution equation. Specifically, water level observation data available on the Internet is collected by web scraping. As well, the graph structure is defined based on numerical river information published by Ministry of Land, Infrastructure, Transport and Tourism (MILT) of Japan and the graph is used to construct GFBs for analyzing and synthesizing the water level data. GFBs work in combination with a sparse approximation algorithm for feature extraction of water level distribution. The features are exploited to derive the time-evolution equation through the extended DMD (EDMD) framework. The time-evolution equation is applied to predict river water level distribution. In order to verify the significance of the proposed method, the river water level prediction is conducted for real web-scraped data. The performance evaluation shows the superiority to the normal DMD approach.

The ultimate goal of this research is to develop the design guidelines for a mobile robot that allows people to easily recognize its movement direction in the scenes where a person and a robot pass each other. In general, the design of a robot coexisting with people often incorporates the human movement and the sensory characteristics in order to let the robot behave like a human. This study also adopts this methodology and take into account with the human behavioral characteristics so that a person can easily recognize the robot's moving direction. Therefore, in this study, we clarify the characteristics of human appearance and movement in the period immediately before passing each other in case of person to person. By adopting its characteristics, we propose the design guidelines of the mobile robot in the point of view of robot shape and movement. In this paper, we performed the experiments ((1) Which body parts of oncoming pedestrian are often gazed by the other pedestrian? (2) Phase difference of yaw angle of the shoulder and the loins, (3) Impression evaluation to derive the most human like shape) to determine robot shape. As a result, we are elucidated as follows. Results of (1): pedestrians are considered to determine their own avoidance direction by seeing the movement of the upper body of the oncoming pedestrian. (2) The rotation of loins shows phase lead to shoulder in case of angle of walking direction change is shallow, on the contrary the shoulder is phase lead to the loins. (3) The most human like shape is that the width of the upper part of the robot, the width of the middle part and the height of the middle part are intermediate, and the width of the lower part is slightly wide.

We have developed a novel optical fiber sensor network system with hetero-core spliced fiber optic sensors. Our previous work demonstrated that the system can play a dual role as a gigabit-class communication network and a sensing system when a limited number of sensors are used. However, optical devices are generally expensive. Thus, in order to use the devices efficiently, it is required that the system allow more sensors to be inserted into a fiber line. In this paper, we propose an enhanced system that can improve sensor multiplicity. The system utilizes a new type of sensor module that generates optical loss only for a short time. We analyze the relationship between sensor multiplicity and communication quality when using such sensor modules, and present results of the performance of this enhanced system in simulated experiments.