Inflammation is an anatomical and physiological response underlying a variety of diseases, and the detection of tissue lesions without subjective symptoms is considered important for preventing of chronic inflammation. For this purpose, it is expected to be useful that recently developed technologies for obtaining single-cell transcriptome data and analytical techniques such as MAGIC and Monocle. In this study, we propose a gene correlation network based on single-cell transcriptome data as a novel analytical approach. Specifically, wep ropose a method for constructing a gene correlation network using MAGIC, and visualization and ranking methods for analyzing its time-series changes. In order to confirm the usefulness of the proposed methods, we conducted experiments using the time-sequence data of single-cell transcriptome obtained from mice induced pulmonary lung fibrosis bye xposure to silica particles. We observed significant changes in the gene correlation network consisting of “negative edges” with the progression of the disease state, as well as that characteristic gene groups with large fluctuations in ranking play important roles in the early stage of the disease.

Our recent proposed notion Geo-Graph-Indistinguishability (GeoGI) is a theoretically rigorous location privacy notion which extends differential privacy to the setting of location privacy on a road network modeled by a graph. GeoGI is practically more useful for protecting location privacy in location-based services (LBSs) such as UBER than a previously proposed Geo-Indistinguishability (Geo-I), which is a popular notion of location privacy defined on the Euclidean plane. We also proposed a Graph-Exponential Mechanism (GEM) based on GeoGI. In this paper, we analyze the utility and privacy of GEM. This analysis makes following two contributions to build a user-friendly location perturbation system. First, our proposed system makes a privacy parameter easy to interpret. The privacy parameter of GEM is not interpretable for ordinary people. By converting to the interpretable value, a user can choose his/her preferred privacy level and utility. Second, we formalize an optimization problem to find the optimal range of GEM according to an arbitrary LBS w.r.t the utility-privacy trade-off for a certain user. Experimental results with real-world data show that our optimization improves the utility-privacy trade-off.

We address a new multi-user routing problem: mutually beneficial confluent routing (MCR). In the MCR, every user has his/her own source and destination; confluences of user routes occur so that users can mutually benefit from travelling together on the confluences. The idea of gaining benefit from travelling together is valuable in various practical applications, such as ride sharing, delivery routing, and pedestrian navigation. We formulate the MCR as a new combinatorial optimization problem on road networks to find optimum routes. As the main contributions of this paper [1], we propose exact and efficient algorithms for the MCR for the setting of two or three users. The setting is reasonable for various practical applications. We have the following two difficulties in solving the MCR: (D1) enumerating patterns of how the users meet and part in the MCR; and (D2) searching optimal meeting and parting nodes from numerous nodes. To overcome the difficulty (D1), we ascertain “confluence patterns” of the optimal solutions. Regarding (D2), we propose several efficient node pruning techniques. Experiments done on large scale road networks reveal that our algorithms are sufficiently efficient.

The recent increase in attention to ride-sharing applications demonstrates the importance of routing algorithms for multiple users who obtain benefits from confluence, that is, traveling together on all or part of their routes. We propose novel and flexible formulation of routing optimization for multiple users who have their respective sources and a single common destination. The formulation is general enough to express each user's benefit (or cost) of confluence for every combination of users. Hence, the formulation can represent a wide range of applications and subsumes almost all formulations proposed in literature. We establish an efficient exact method for the formulation. Interestingly, we found well-known Dreyfus-Wagner Algorithm for the Minimum Steiner Tree Problem (MSTP) is extensible for ours, although our formulation is much harder than the MSTP. Our experimental results obtained on large-scale road networks reveal that our method is efficient in practical settings.

We investigate a new multi-user routing problem: mutually beneficial confluent routing (MCR). In the MCR, every user has his/her own source and destination; confluences of user routes occur so that users can mutually benefit from travelling together on the confluences. The idea of gaining benefit from travelling together is valuable in various practical applications, such as ride sharing, delivery routing, and pedestrian navigation. We formulate the MCR as a new combinatorial optimization problem on road networks. The MCR is more general and complex than single vehicle routing problems, ride-sharing problems, and the Steiner tree problem. We propose exact and efficient algorithms for the MCR for the setting of two or three users. The setting is reasonable for various practical applications. The key ideas of our algorithms are to use “confluence patterns” of the optimal solutions and exploit the properties of geometric graphs. Experimental results obtained on large scale road networks reveal that our algorithms are sufficiently efficient.

Walking is an easy exercise for maintaining health, although there are several environmental factors, such as road congestion or slopes, which are stressful for pedestrians, especially elders. Therefore, a pedestrian navigation system is required to recommend a less stressful route for a user by estimating such environmental factors. Towards realizing such a system, we firstly collect vital sign, spatiotemporal and environmental data from pedestrians, and secondly analyze the relation between them, and then propose models for estimating the personal stress of a pedestrian. As our vital sign data, we calculate the stress of each pedestrian from the R-R interval data obtained by a heart rate sensor. Our spatiotemporal data includes the walking speed of pedestrians and road grade calculated from GPS sensor data. We measure the degree of road congestion, as our environmental data, using a laser range finder. Our analysis reveals that the road congestion is an important factor affecting stress in walking. The proposed models based on neural networks employ the road congestion in addition to the vital sign and spatiotemporal data. Experimental results indicate that the road congestion degree is effective and the stress of pedestrians estimated by our models is similar to the real stress.

In image retrieval systems with interfaces of user relevance feedback, different users label different instances based on different image search results they prefer. In our previous work, we proposed a multi-model relevance feedback scheme for social image retrieval, which allows users to label relevance feedback information on different media modalities. These relevance feedback instances contain various potential semantics information related to users' image targets. In this work-in-progress paper, we analyze various cases of user multi-model relevance feedback selections and their potential semantics to the targets, and then propose an idea of categorization for them. In future work, we will improve our approach to meets users' requirements leveraging these knowledge.

We focus on measuring relationships between pairs of objects in Wikipedia whose pages can be regarded as individual objects. Two kinds of relationships between two objects exist: in Wikipedia, an explicit relationship is represented by a single link between the two pages for the objects, and an implicit relationship is represented by a link structure containing the two pages. Some of the previously proposed methods for measuring relationships are cohesion-based methods, which underestimate objects having high degrees, although such objects could be important in constituting relationships in Wikipedia. The other methods are inadequate for measuring implicit relationships because they use only one or two of the following three important factors: distance, connectivity, and cocitation. We propose a new method using a generalized maximum flow which reflects all the three factors and does not underestimate objects having high degree. We confirm through experiments that our method can measure the strength of a relationship more appropriately than these previously proposed methods do. Another remarkable aspect of our method is mining elucidatory objects, that is, objects constituting a relationship. We explain that mining elucidatory objects would open a novel way to deeply understand a relationship.

For understanding human activity, a useful approach is to rank people according to the strength of their relations to a specified person. Similarly, rankings of objects based on such a relation are used in several fields. Several methods have been proposed for computing the strength of a relation between two objects. These methods do not present a reason why an object has a stronger relation to a specified object than another object has. Therefore, it is difficult for a user to understand a ranking obtained using these methods. On the other hand, the authors recently proposed a method for computing the strength of a relation through mining objects to elucidate the relation. We propose a ranking tool based on this method to afford a better understanding of a ranking. Our ranking tool has the following three characteristics for understanding a ranking: (1) it visualizes a relation by displaying objects that are important for elucidating the relation; (2) it classifies ranked objects into several groups, e.g., the two groups of "petroleum exporting countries'' and "petroleum consuming countries'' for ranking countries based on their respective relations to petroleum; and (3) it visualizes a reason explaining why an object has a stronger relation to a specified object than another object has. We explain novel ideas used in our ranking tool for understanding a ranking based on relations, and claim how effective the features of our tool are by presenting examples.

In recent years, several information retrieval methods using information about Web-links have been developed, such as HITS and trawling. In order to analyze Web-links dividing into links inside each Web site (local-links) and links between Web sites (global-links)for information retrieval, a proper model of the Web site is required. In existing research, a Web server is used as a model of the Web site. This idea works relatively well when a Web site corresponds to a server, as is the case for public Web sites, but works poorly when multiple Web sites correspond to a server, as is the case for private Web sites on rental Web servers. We propose a new model of the Web site, "directory-based site", to handle typical private sites, and a method to identify them using information about the URL and Web-links. We verify the method can approximately identify, at a rate of 66% of over 110,000 servers, whether each server has multiple directory-based sites or not, and extract over 500,000 directory-based sites and 4 million global-links by computational experiments using jp-domain URLs and Web-link data contains over 23 million URLs and 100 million Web-links, collected from July to August 2000, by Toyoda and Kitsuregawa. We also propose a new framework of Web-link based information retrieval that uses directory-based sites and global-links instead of Web pages and whole Web-links respectively, and examine the effectiveness of our framework by comparing a result of trawling on our framework to one on the existing framework.