Information and System Group

December 2014

1  2014-12-08 K-SVD: An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation Haeung
 Using an overcomplete dictionary that contains prototype of signal-atoms, signals are described by sparse linear combinations of these atoms. In this paper, authors introduce a novel algorithm for adapting dictionaries in order to achieve sparse representations—the k-SVD algorithm. This algorithm repeats sparse coding with current dictionary and updating the dictionary to better fit the data. And the authors analyze this algorithm and demonstrate its results.
2  2014-12-22 Fronthaul Compression for Cloud Radio Access Networks Zafar  (doc)
Cloud Radio Access Networks (C-RANs) are novel next-generation wireless cellular network architectures where the baseband processing is shifted from the base station (BS) to the control unit (CU) in the “cloud” of the operator. In C-RANs, unlike the current cellular systems, the functionalities needed to process the IQ samples of the radio signals received/transmitted by the RUs are not implemented in the RUs but in the CU within the cloud of the core network. The BSs which function as radio units (RUs) are connected to the CU via fronthaul links. The fronthaul links carry information to the CU from the RU and vice versa in the form of quantized in-phase and quadrature (IQ) samples. However, the current solutions for standardization efforts use conventional scalar quantization techniques which impose a bottleneck to the system performance due to limitations in the fronthaul capacity. Due the large amount of IQ data generated by the network and limited capacity of the fronthaul links, compression prior to transmission on the fronthaul is very necessary and is of significant importance in the design of next generation wireless cellular networks. This article provides a summary of the recent work in this area with emphasis on the advanced signal processing solutions.
3  2014-12-29 Information-theoretically Optimal Sparse PCA Jehyuk  (pdf)
 Sparse Principal Component Analysis (PCA) is a
dimensionality reduction technique wherein one seeks a lowrank representation of a data matrix with additional sparsity
constraints on the obtained representation. We consider two
probabilistic formulations of sparse PCA: a spiked Wigner and
spiked Wishart (or spiked covariance) model. We analyze an
Approximate Message Passing (AMP) algorithm to estimate the
underlying signal and show, in the high dimensional limit, that
the AMP estimates are information-theoretically optimal. As an
immediate corollary, our results demonstrate that the posterior
expectation of the underlying signal, which is often intractable to
compute, can be obtained using a polynomial-time scheme. Our
results also effectively provide a single-letter characterization of
the sparse PCA problem.