Nikolai Nefedov

Clustering on Multi-Layer Graphs via Subspace Analysis on Grassmann Manifolds

Relationships between entities in datasets are often of multiple nature, like geographical distance, social relationships, or common interests among people in a social network, for example. This information can naturally be modeled by a set of weighted and undirected graphs that form a global multi-layer graph, where the common vertex set represents the entities and the edges on different layers capture the similarities of the entities in term of the different modalities. In this paper, we address the problem of analyzing multi-layer graphs and propose methods for clustering the vertices by efficiently merging the information provided by the multiple modalities. To this end, we propose to combine the characteristics of individual graph layers using tools from subspace analysis on a Grassmann manifold. The resulting combination can then be viewed as a low dimensional representation of the original data which preserves the most important information from diverse relationships between entities. As an illustrative application of our framework, we use our algorithm in clustering methods and test its performance on several synthetic and real world datasets where it is shown to be superior to baseline schemes and competitive to state-of-the-art techniques. Our generic framework further extends to numerous analysis and learning problems that involve different types of information on graphs.

Clustering With Multi-Layer Graphs: A Spectral Perspective

Observational data usually comes with a multimodal nature, which means that it can be naturally represented by a multi-layer graph whose layers share the same set of vertices (objects) with different edges (pairwise relationships). In this paper, we address the problem of combining different layers of the multi-layer graph for an improved clustering of the vertices compared to using layers independently. We propose two novel methods, which are based on a joint matrix factorization and a graph regularization framework respectively, to efficiently combine the spectrum of the multiple graph layers, namely the eigenvectors of the graph Laplacian matrices. In each case, the resulting combination, which we call a “joint spectrum” of multiple layers, is used for clustering the vertices. We evaluate our approaches by experiments with several real world social network datasets. Results demonstrate the superior or competitive performance of the proposed methods compared to state-of-the-art techniques and common baseline methods, such as co-regularization and summation of information from individual graphs.

Significance of Nanotechnology for Future Wireless Devices and Communications

This paper reviews the expected wide and profound impact of nanotechnology for future wireless devices and communication technologies.

Iterative data detection and channel estimation for advanced TDMA systems

This letter presents a new receiver for Q-ary transmission, where all receiver blocks are embedded in an iterative structure. Packet data transmission in Global Systems for Mobile communications (GSM) and Enhanced Data rates for Global Evolution (EDGE) are considered as examples. A low-complexity soft-in-soft-out detector for EDGE is introduced and its modification suitable for iterative detection is derived. Application of iterative detection and channel estimation techniques in GSM/EDGE shows a significant performance enhancement. Additional improvement may be obtained if the iterative processing is applied to packet retransmission schemes.

Iterative channel estimation for GPRS

In this paper we consider iterative estimation and equalization techniques and present a simple method of updating channel estimates that includes decoder outputs into the iteration process. To clarify performance-complexity trade-off we evaluated iterative estimation and turbo equalization techniques in the General Packet Radio System (GPRS). It is found that turbo estimation is more beneficial for GPRS, showing 1 dB gain after one channel estimate update, while turbo equalization rounds provide only a slight improvement on the top of that.

Evaluation of low complexity concatenated codes for high data rate transmission

In this paper we compare low complexity turbo-like codes formed by differential encoders. A unified coding scheme combining turbo codes and parallel concatenated zigzag codes (PCZZ) and its multi-dimensional extension for packet re-transmission schemes in HSDPA/WCDMA is proposed. It is shown that low complexity PCZZ outperforms WCDMA turbo codes at high coding rates. Modified zigzag decoder and suggested simple stopping rule for PCZZ iterative decoding brings further complexity reduction without visible performance loss.

Generative Markov models for discrete channel modelling

To speedup the simulations aimed at different error control and interleaving schemes evaluation there is a need for generative models of the radio layer communication link. A method of modelling discrete channels with hard and soft decision outputs is suggested. This method is based on nested hidden Markov model (HMM) and implemented as COSSAP library modules. The verification made by simulations of GSM speech traffic confirmed the validity of suggested approach.

Evaluation of potential transmit diversity schemes with iterative receivers in EDGE

In this paper we compare conventional and iterative receiver structures for transmit diversity schemes for frequency selective fading channels. As a practical example we consider an EDGE system with two transmitting and one receiving antennas. In particular, delay diversity (DD) and space-time block codes (STBC) for multipath channels are considered. It is shown that STBC in EDGE brings up to 3 dB gain, which is by 1-2 dB more than the gain provided by the DD. However, the SBTC diversity gain may not be fully utilized due to degradation of channel estimation accuracy. To improve the accuracy of channel estimates and the overall performance we consider receivers with iterative (turbo) estimation/detection. It is shown that iterative data processing in EDGE is more beneficial for STBC than for the DD.

Discrete channel models for wireless data services

The evaluation of error control strategies for different wireless services implies the existence of models which can describe physical and data-link layers as a whole. In this paper a method of modelling of discrete channels (DC) is presented. This method is based on nested hidden Markov models and takes into consideration hard and soft decision outputs. It is shown that realistic DC models built for GSM radio channels (at bit level as well as at TDMA frame level) may have a relatively large number of channel states. DC model parameters for ARQ and hybrid FEC-ARQ schemes are presented.

Iterative receiver concept for TDMA packet data systems

In this paper we present an iterative receiver concept that improves the radio link performance of TDMA mobile communication systems. We consider sub-optimal receiver structures comprising of channel estimator, detector and channel decoder, where the performance is improved by iterative data processing among the receiver blocks. As a practical example, we consider packet data transmission in GSM and Enhanced Data rates for Global Evolution (EDGE). Low-complexity soft-in-soft-out (SISO) equalizers for EDGE are introduced and its modifications suitable for iterative detection in EDGE are derived. Application of iterative detection and channel estimation techniques in GSM/EDGE shows a significant performance improvement. Furthermore, we show that retransmission schemes specified for EDGE also benefit from iterative data processing. Copyright