Vladimir Krylov

Scalable distributed algorithm for approximate nearest neighbor search problem in high dimensional general metric spaces

We propose a novel approach for solving the approximate nearest neighbor search problem in arbitrary metric spaces. The distinctive feature of our approach is that we can incrementally build a non-hierarchical distributed structure for given metric space data with a logarithmic complexity scaling on the size of the structure and adjustable accuracy probabilistic nearest neighbor queries. The structure is based on a small world graph with vertices corresponding to the stored elements, edges for links between them and the greedy algorithm as base algorithm for searching. Both search and addition algorithms require only local information from the structure. The performed simulation for data in the Euclidian space shows that the structure built using the proposed algorithm has navigable small world properties with logarithmic search complexity at fixed accuracy and has weak (power law) scalability with the dimensionality of the stored data.

Toward the power InterGrid

The paper is focused on the new approach to electric power distribution. Following the analogy of telecommunications network and power grids evolution a power packetizing technology is proposed. The new concept of the Packet Power Grid (PPG) is proposed as a set of different power sources and power consumers joined by power lines and electronic switches controlled by a supporting information computer network. Different electrical sensors form an important part of the PPG as well. The paper overviews the PPG as the next power grid evolution step after the Smart Grid. The PPG approach solves a number of problems of open concurrent power market development e.g. highly-efficient energy usage and numerous small sources adoption. The growing Packet Power Grid can play the role of the global power highway the same way as the Internet plays the role of the global information highway.

Metrized Small World Approach for Nearest Neighbor Search

In different areas attempts are made to organize data into multi-linked structures which are well suited for information search, in particular the nearest neighbor search where the result data items are metrically close to a given data item. These structures often take the form of trees (M-Tree, cover tree, KD-tree, GNAT) or networks (M-Chord, VoroNet, RayNet) built over a set of data items. In this paper we give the regular approach to the construction of links between data items which provides logarithmical time complexity of the nearest neighbor search in the structure. According to this approach, data items are organized into an undirected graph with Small World properties, which ensure the existence of a short path between any two data items regardless of the graph size. We propose different construction and search algorithms depending on the properties of the metric which determines the proximity of data items. The types of metric we consider are abstract metric and ordered metric. Further we extend the ordered metric approach to compound data items in the form of attribute-value pair sets to enable inclusion search by an arbitrary subset of attribute-value pairs. Finally we provide simulation results for the structure with compound data items.

Odor Space Navigation Using Multisensory E-Nose

A spatial odor distribution in an environment can be used for navigation, goal search, localization and mapping, like by video, ultrasonic, temperature and other sensors. Modern e-noses can perform the selective detection of different gases with an extremely low concentration but the source localization algorithms of a selected gas against the background of other odors are still underinvestigated. This paper studies an odor field representation in terms of an e-nose based on an array of low-selective sensors. Using a simulation model, we show how the vector measurements of a field of several odor sources can be processed to navigate for reaching a selected odor source. In addition, we demonstrate that the source having a high level of odor intensity can interfere with the search of another odor source of a low intensity. The well-known class of matching receivers does not solve this problem. However, a solution can be obtained by distributed measurements. As shown below, the spatial structure of an odor field allows to implement vector selection. Using deep learning machines, we may reach a high resolution of odor sources in the space. Our future research will be focused on augmented odor reality and autonomous mobile e-nose (e-dog) design.

The convoluted multiaddress networking architecture principles and application

To increase robustness of network nodes and their communication sessions, we propose convoluted multiaddress networking architecture. This approach prevents malicious packets from getting into the incoming traffic of a network terminal. Usually, traffic analyzers and filtering solutions should be installed in the network to isolate a victim node from packet streams created by malefactor terminals. Our network security technique is built on a different approach. The principles of convoluted multiaddress networks are based on the idea that we can protect nodes by hiding their network location from illegitimate clients. In our study, we show how to create dynamic addressing policies for preventing DDoS attacks and traffic eavesdropping. These policies randomize address space and communication data streams, therefore a malefactor cannot acquire access to data streams or destination terminals. In this paper, we discuss IP Fast Hopping, an application of convoluted multiaddress networking in TCP/IP networks. We consider basic implementation of this architecture, its major practical constraints and initial experimental results. The presented approach aims to ensure security of future generation communication technologies. In this study, we suggest Thing Lakes architecture of the Internet of Things, which is based on IP Fast Hopping approach and intended to protect the IoT environment against several major security issues in such networks.