Publications (208)181.47 Total impact
 [Show abstract] [Hide abstract]
ABSTRACT: Under the paradigm of caching, partial data is delivered before the actual requests of the users are known. In this paper, this problem is modeled as a canonical distributed source coding problem with side information, where the side information represents the users' requests. For the singleuser case, a singleletter characterization of the optimal rate region is established, and for several important special cases, closedform solutions are given, including the scenario of uniformly distributed user requests. In this case, it is shown that the optimal caching strategy is closely related to total correlation and Wyner's common information. For the twouser case, five representative subproblems are considered, which draw connections to existing source coding problems in the literature: the GrayWyner system, distributed successive refinement, and the Kaspi/HeegardBerger problem.  [Show abstract] [Hide abstract]
ABSTRACT: In this paper we study the data exchange problem where a set of users is interested in gaining access to a common file, but where each has only partial knowledge about it as sideinformation. Assuming that the file is broken into packets, the sideinformation considered is in the form of linear combinations of the file packets. Given that the collective information of all the users is sufficient to allow recovery of the entire file, the goal is for each user to gain access to the file while minimizing some communication cost. We assume that users can communicate over a noiseless broadcast channel, and that the communication cost is a sum of each user's cost function over the number of bits it transmits. For instance, the communication cost could simply be the total number of bits that needs to be transmitted. In the most general case studied in this paper, each user can have any arbitrary convex cost function. We provide deterministic, polynomialtime algorithms (in the number of users and packets) which find an optimal communication scheme that minimizes the communication cost. To further lower the complexity, we also propose a simple randomized algorithm inspired by our deterministic algorithm which is based on a random linear network coding scheme.  [Show abstract] [Hide abstract]
ABSTRACT: ComputeandForward is an emerging technique to deal with interference. It allows the receiver to decode a suitably chosen integer linear combination of the transmitted messages. The integer coefficients should be adapted to the channel fading state. Optimizing these coefficients is a Shortest Lattice Vector (SLV) problem. In general, the SLV problem is known to be prohibitively complex. In this paper, we show that the particular SLV instance resulting from the ComputeandForward problem can be solved in low polynomial complexity and give an explicit deterministic algorithm that is guaranteed to find the optimal solution.  [Show abstract] [Hide abstract]
ABSTRACT: Lattice codes used under the ComputeandForward paradigm suggest an alternative strategy for the standard Gaussian multipleaccess channel (MAC): The receiver successively decodes integer linear combinations of the messages until it can invert and recover all messages. In this paper, this strategy is developed and analyzed. For the twouser MAC, it is shown that without timesharing, the entire capacity region can be attained with a singleuser decoder as soon as the signaltonoise ratios are above $1+\sqrt{2}$. A partial analysis is given for more than two users. Lastly the strategy is extended to the socalled dirty MAC where two interfering signals are known noncausally to the two transmitters in a distributed fashion. Our scheme extends the previously known results and gives new achievable rate regions.  [Show abstract] [Hide abstract]
ABSTRACT: ComputeandForward (CF), also known as reliable physical layer network coding, is a technique that provides the possibility of exploiting the features of broadcast and superposition in wireless networks. It has been shown that the throughput for multiple unicast traffic can be significantly boosted by CF. In this letter, the limit of such improvement is investigated by comparing the performance of CF with the traditional routingbased transmission schemes. For networks characterized by local interference and halfduplex constraints, it is proven that the throughput gain of CF over traditional routing, expressed by an improvement factor, is upper bounded by $3K$, where $K$ is the number of unicast sessions. Furthermore, a class of networks is presented for which an improvement by a factor of $K/2$ is feasible by applying CF. Hence, the throughput gain of CF is at most on the order of $K$ for any network, and a gain in that order is indeed achievable for some networks.IEEE Communications Letters 07/2014; 18(7):11111114. DOI:10.1109/LCOMM.2014.2323242 · 1.46 Impact Factor 
Conference Paper: On distributed successive refinement with lossless recovery
[Show abstract] [Hide abstract]
ABSTRACT: The problem of successive refinement in distributed source coding and in joint sourcechannel coding is considered. The emphasis is placed on the case where the sources have to be recovered losslessly in the second stage. In distributed source coding, it is shown that all sources are successively refinable in sum rate, with respect to any (joint) distortion measure in the first stage. In joint sourcechannel coding, the sources are assumed independent and only a (per letter) function is to be recovered losslessly in the first stage. For a class of multiple access channels, it is shown that all sources are successively refinable with respect to a class of linear functions. Finally, when the sources have equal entropy, a simple sufficient condition of successive refinability is provided for partially invertible functions.2014 IEEE International Symposium on Information Theory (ISIT); 06/2014  [Show abstract] [Hide abstract]
ABSTRACT: Lattice codes are applied to the twouser Gaussian multiple access channel (MAC) combined with a modified computeandforward transmitting scheme. It is shown that noncorner points on the boundary of the capacity region can be achieved by decoding two integer sums of the codewords, which can be viewed as a generalization of the wellknown successive cancellation decoding. A similar idea is then applied to the socalled dirty MAC where two interfering signals are known noncausally to the two transmitters in a distributed fashion. Our scheme recovers previously known results and gives new achievable rate regions. The proposed scheme can be extended to the case with more than two users.2014 IEEE International Symposium on Information Theory (ISIT); 06/2014  [Show abstract] [Hide abstract]
ABSTRACT: The Shortest Lattice Vector (SLV) problem is in general hard to solve, except for special cases (such as root lattices and lattices for which an obtuse superbase is known). In this paper, we present a new class of SLV problems that can be solved efficiently. Specifically, if for an $n$dimensional lattice, a Gram matrix is known that can be written as the difference of a diagonal matrix and a positive semidefinite matrix of rank $k$ (for some constant $k$), we show that the SLV problem can be reduced to a $k$dimensional optimization problem with countably many candidate points. Moreover, we show that the number of candidate points is bounded by a polynomial function of the ratio of the smallest diagonal element and the smallest eigenvalue of the Gram matrix. Hence, as long as this ratio is upper bounded by a polynomial function of $n$, the corresponding SLV problem can be solved in polynomial complexity. Our investigations are motivated by the emergence of such lattices in the field of Network Information Theory. Further applications may exist in other areas.  [Show abstract] [Hide abstract]
ABSTRACT: A new achievable rate region is given for the Gaussian cognitive manytoone interference channel. The proposed coding scheme is based on lattice codes and incorporates all key ingredients from conventional coding schemes. Using the idea of computeandforward to decode sums of codewords, our scheme improves considerably upon the conventional coding schemes which treat interference as noise or decode messages simultaneously. Roughly speaking, the novel scheme asks the decoder to decode "just enough" interfering signals and use it for extracting the desired information. Our strategy extends directly to the usual manytoone interference channels without cognitive messages. For some simple channel settings, the proposed scheme gives constant gap or capacity results which are independent of the number of users in the system.IEEE Transactions on Information Theory 04/2014; 61(3). DOI:10.1109/TIT.2015.2394782 · 2.65 Impact Factor  [Show abstract] [Hide abstract]
ABSTRACT: Objective. Intracortical brainmachine interfaces (BMIs) have predominantly utilized spike activity as the control signal. However, an increasing number of studies have shown the utility of local field potentials (LFPs) for decoding motor related signals. Currently, it is unclear how well different LFP frequencies can serve as features for continuous, closedloop BMI control. Approach. We demonstrate 2D continuous LFPbased BMI control using closedloop decoder adaptation, which adapts decoder parameters to subjectspecific LFP feature modulations during BMI control. We trained two macaque monkeys to control a 2D cursor in a centerout task by modulating LFP power in the 0150 Hz range. Main results. While both monkeys attained control, they used different strategies involving different frequency bands. One monkey primarily utilized the lowfrequency spectrum (080 Hz), which was highly correlated between channels, and obtained proficient performance even with a single channel. In contrast, the other monkey relied more on higher frequencies (80150 Hz), which were less correlated between channels, and had greater difficulty with control as the number of channels decreased. We then restricted the monkeys to use only various subranges (040, 4080, and 80150 Hz) of the 0150 Hz band. Interestingly, although both monkeys performed better with some subranges than others, they were able to achieve BMI control with all subranges after decoder adaptation, demonstrating broad flexibility in the frequencies that could potentially be used for LFPbased BMI control. Significance. Overall, our results demonstrate proficient, continuous BMI control using LFPs and provide insight into the subjectspecific spectral patterns of LFP activity modulated during control.Journal of Neural Engineering 02/2014; 11(2):026002. DOI:10.1088/17412560/11/2/026002 · 3.42 Impact Factor  [Show abstract] [Hide abstract]
ABSTRACT: We present a modified computeandforward scheme which utilizes Channel State Information at the Transmitters (CSIT) in a natural way. The modified scheme allows different users to have different coding rates, and use CSIT to achieve larger rate region. This idea is applicable to all systems which use the computeandforward technique and can be arbitrarily better than the regular scheme in some settings.  [Show abstract] [Hide abstract]
ABSTRACT: Function computation of arbitrarily correlated discrete sources over Gaussian networks with orthogonal components is studied. Two classes of functions are considered: the arithmetic sum function and the type function. The arithmetic sum function in this paper is defined as a set of multiple weighted arithmetic sums, which includes averaging of the sources and estimating each of the sources as special cases. The type or frequency histogram function counts the number of occurrences of each argument, which yields many important statistics such as mean, variance, maximum, minimum, median, and so on. The proposed computation coding first abstracts Gaussian networks into the corresponding modulo sum multipleaccess channels via nested lattice codes and linear network coding and then computes the desired function by using linear SlepianWolf source coding. For orthogonal Gaussian networks (with no broadcast and multipleaccess components), the computation capacity is characterized for a class of networks. For Gaussian networks with multipleaccess components (but no broadcast), an approximate computation capacity is characterized for a class of networks.10/2013; 60(12). DOI:10.1109/ISIT.2013.6620604 
Article: Interactive Computation of TypeThreshold Functions in Collocated BroadcastSuperposition Networks
[Show abstract] [Hide abstract]
ABSTRACT: In wireless sensor networks, various applications involve learning one or multiple functions of the measurements observed by sensors, rather than the measurements themselves. This paper focuses on typethreshold functions, e.g., the maximum and indicator functions. Previous work studied this problem under the collocated collision network model and showed that under many probabilistic models for the measurements, the achievable computation rates converge to zero as the number of sensors increases. This paper considers two network models reflecting both the broadcast and superposition properties of wireless channels: the collocated linear finite field network and the collocated Gaussian network. A general multiround coding scheme exploiting not only the broadcast property but particularly also the superposition property of the networks is developed. Through careful scheduling of concurrent transmissions to reduce redundancy, it is shown that given any independent measurement distribution, all typethreshold functions can be computed reliably with a nonvanishing rate in the collocated Gaussian network, even if the number of sensors tends to infinity.  [Show abstract] [Hide abstract]
ABSTRACT: This paper explores the problem of feedback coding for a channel whose output is simultaneously used for two purposes: it is decoded to establish reliable communication, and it is used to control a dynamical system. In general, there is a tradeoff between the rate of communication and the accuracy of control. An intuitive communication and control strategy is analyzed and shown to be optimal in several cases of interest. Using methods from stochastic control, a corresponding upper bound is derived on the rate for a given cost, and this bound can be applied to find the capacity for an interesting class of channels. Under certain regularity conditions, the capacity has a simple characterization, and a series of examples is provided to demonstrate how it can be calculated.IEEE Transactions on Information Theory 10/2013; 59(10):62436257. DOI:10.1109/TIT.2013.2272035 · 2.65 Impact Factor  [Show abstract] [Hide abstract]
ABSTRACT: Recent progress in brainmachine interfaces (BMIs) has shown tremendous improvements in task complexity and degree of control. In particular, closedloop decoder adaptation (CLDA) has emerged as an effective paradigm for both improving and maintaining the performance of BMI systems. Here, we demonstrate the first reported use of a CLDA algorithm to rapidly achieve highperformance control of a BMI based on local field potentials (LFPs). We trained a nonhuman primate to control a 2D computer cursor by modulating LFP activity to perform a centerout reaching task, while applying CLDA to adaptively update the decoder. We show that the subject is quickly able to readily reach and hold at all 8 targets with an average success rate of 74% ± 7% (sustained peak rate of 85%), with rapid convergence in the decoder parameters. Moreover, the subject is able to maintain high performance across 4 days with minimal adaptations to the decoder. Our results indicate that CLDA can be used to facilitate LFPbased BMI systems, allowing for both rapid improvement and maintenance of performance.Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference 07/2013; 2013:285288. DOI:10.1109/EMBC.2013.6609493  [Show abstract] [Hide abstract]
ABSTRACT: It is plausible to hypothesize that the spiking responses of certain neurons represent functions of the spiking signals of other neurons. A natural ensuing question concerns how to use experimental data to infer what kind of a function is being computed. Modelbased approaches typically require assumptions on how information is represented. By contrast, information measures are sensitive only to relative behavior: information is unchanged by applying arbitrary invertible transformations to the involved random variables. This paper develops an approach based on the information bottleneck method that attempts to find such functional relationships in a neuron population. Specifically, the information bottleneck method is used to provide appropriate compact representations which can then be parsed to infer functional relationships. In the present paper, the parsing step is specialized to the case of remappedlinear functions. The approach is validated on artificial data and then applied to recordings from the motor cortex of a macaque monkey performing an armreaching task. Functional relationships are identified and shown to exhibit some degree of persistence across multiple trials of the same experiment.Entropy 05/2013; 5(5). DOI:10.3390/e15051587 · 1.56 Impact Factor  [Show abstract] [Hide abstract]
ABSTRACT: Previous research has shown that postnatal exposure to simple, synthetic sounds can affect the sound representation in the auditory cortex as reflected by changes in the tonotopic map or other relatively simple tuning properties, such as AM tuning. However, their functional implications for neural processing in the generation of ethologicallybased perception remain unexplored. Here we examined the effects of noiserearing and social isolation on the neural processing of communication sounds such as speciesspecific song, in the primary auditory cortex analog of adult zebra finches. Our electrophysiological recordings reveal that neural tuning to simple frequencybased synthetic sounds is initially established in all the laminae independent of patterned acoustic experience; however, we provide the first evidence that early exposure to patterned sound statistics, such as those found in native sounds, is required for the subsequent emergence of neural selectivity for complex vocalizations and for shaping neural spiking precision in superficial and deep cortical laminae, and for creating efficient neural representations of song and a less redundant ensemble code in all the laminae. Our study also provides the first causal evidence for 'sparse coding', such that when the statistics of the stimuli were changed during rearing, as in noiserearing, that the sparse or optimal representation for speciesspecific vocalizations disappeared. Taken together, these results imply that a layerspecific differential development of the auditory cortex requires patterned acoustic input, and a specialized and robust sensory representation of complex communication sounds in the auditory cortex requires a rich acoustic and social environment.PLoS ONE 04/2013; 8(4):e61417. DOI:10.1371/journal.pone.0061417 · 3.23 Impact Factor 
Article: Polar Codes For Broadcast Channels
[Show abstract] [Hide abstract]
ABSTRACT: Polar codes are introduced for discrete memoryless broadcast channels. For $m$user deterministic broadcast channels, polarization is applied to map uniformly random message bits from $m$ independent messages to one codeword while satisfying broadcast constraints. The polarizationbased codes achieve rates on the boundary of the privatemessage capacity region. For twouser noisy broadcast channels, polar implementations are presented for two informationtheoretic schemes: i) Cover's superposition codes; ii) Marton's codes. Due to the structure of polarization, constraints on the auxiliary and channelinput distributions are identified to ensure proper alignment of polarization indices in the multiuser setting. The codes achieve rates on the capacity boundary of a few classes of broadcast channels (e.g., binaryinput stochastically degraded). The complexity of encoding and decoding is $O(n*log n)$ where $n$ is the block length. In addition, polar code sequences obtain a stretchedexponential decay of $O(2^{n^{\beta}})$ of the average block error probability where $0 < \beta < 0.5$.01/2013; 61(2). DOI:10.1109/ISIT.2013.6620402 
Conference Paper: Network decomposition for function computation
[Show abstract] [Hide abstract]
ABSTRACT: We develop a networkdecomposition framework to provide elementary parallel subnetworks that can constitute an original network without loss of optimality. In our earlier work, a network decomposition is constructed for the AvestimehrDiggaviTse deterministic network which well captures key properties of wireless Gaussian networks. In this work, we apply this decomposition framework to general problem settings where receivers intend to compute functions of the messages generated at transmitters. Depending on functions, these settings include a variety of network problems, ranging from classical communication problems (such as multipleunicast and multicast problems) to function computation problems. For many of these problems, we show that coding separately over the decomposed orthogonal subnetworks provides optimal performances, thus establishing a separation principle.Signal Processing Advances in Wireless Communications (SPAWC), 2013 IEEE 14th Workshop on; 01/2013 
Conference Paper: Interactive function computation
[Show abstract] [Hide abstract]
ABSTRACT: We investigate the role of interaction for computation problem settings where nodes intend to compute functions of the raw messages generated at other nodes. In this work, we make some progress on a more elementary research component: feedback. Specifically we characterize the feedback computing capacity of a twotransmitter tworeceiver linear deterministic network in which both receivers wish to decode a linear function (modulo2 sum) of Bernoulli sources generated at the transmitters. Inspired by the concept of interference alignment and computeandforward, we develop a new achievable scheme called interactive function alignment. A new converse theorem is established that is tighter than cutset based and genieaided bounds. As a consequence of this result, we show that interaction can provide an arbitrarily large gain for computation, as in classical communication settings.Information Theory Proceedings (ISIT), 2013 IEEE International Symposium on; 01/2013
Publication Stats
6k  Citations  
181.47  Total Impact Points  
Top Journals
Institutions

2009–2014

Delft University of Technology
 • Department of Intelligent Systems
 • Faculty of Electrical Engineering, Mathematics and Computer Sciences (EEMCS)
Delft, South Holland, Netherlands


2–2014

University of California, Berkeley
 Department of Electrical Engineering and Computer Sciences
Berkeley, California, United States


2010

University of California, San Diego
San Diego, California, United States 
San Jose State University
 Department of Electrical Engineering
San Jose, CA, United States 
University of Wisconsin, Madison
 Department of Electrical and Computer Engineering
Madison, Wisconsin, United States


2002–2010

École Polytechnique Fédérale de Lausanne
 School of Computer and Communication Sciences
Lausanne, Vaud, Switzerland


2007

Cornell University
 Department of Electrical and Computer Engineering
Итак, New York, United States


2001

Eawag: Das WasserforschungsInstitut des ETHBereichs
Duebendorf, Zurich, Switzerland
