arXiv subjects

Tara Javidi

Publications and source records attributed to Tara Javidi.

86 records · Page 5Linked to original sources

arXiv2015

Adaptive Policies for Scheduling with Reconfiguration Delay: An End-to-End Solution for All-Optical Data Centers

All-optical switching networks have been considered a promising candidate for the next generation data center networks thanks to its scalability in data bandwidth and power efficiency. However, the bufferless nature and the nonzero recon- figuration delay of optical switches remain great challenges in deploying all-optical networks. This paper considers the end-to- end scheduling for all-optical data center networks with no in- network buffer and nonzero reconfiguration delay. A framework is proposed to deal with the nonzero reconfiguration delay. The proposed approach constructs an adaptive variant of any given scheduling policy. It is shown that if a scheduling policy guarantees its schedules to have schedule weights close to the MaxWeight schedule (and thus is throughput optimal in the zero reconfiguration regime), then the throughput optimality is inherited by its adaptive variant (in any nonzero reconfiguration delay regime). As a corollary, a class of adaptive variants of the well known MaxWeight policy is shown to achieve throughput optimality without prior knowledge of the traffic load. Further- more, through numerical simulations, the simplest such policy, namely the Adaptive MaxWeight (AMW), is shown to exhibit better delay performance than all prior work.

cs.NI↗

arXiv2015

Efficient Object Detection for High Resolution Images

Efficient generation of high-quality object proposals is an essential step in state-of-the-art object detection systems based on deep convolutional neural networks (DCNN) features. Current object proposal algorithms are computationally inefficient in processing high resolution images containing small objects, which makes them the bottleneck in object detection systems. In this paper we present effective methods to detect objects for high resolution images. We combine two complementary strategies. The first approach is to predict bounding boxes based on adjacent visual features. The second approach uses high level image features to guide a two-step search process that adaptively focuses on regions that are likely to contain small objects. We extract features required for the two strategies by utilizing a pre-trained DCNN model known as AlexNet. We demonstrate the effectiveness of our algorithm by showing its performance on a high-resolution image subset of the SUN 2012 object detection dataset.

cs.CV↗

arXiv2015

Extrinsic Jensen-Shannon Divergence: Applications to Variable-Length Coding

This paper considers the problem of variable-length coding over a discrete memoryless channel (DMC) with noiseless feedback. The paper provides a stochastic control view of the problem whose solution is analyzed via a newly proposed symmetrized divergence, termed extrinsic Jensen-Shannon (EJS) divergence. It is shown that strictly positive lower bounds on EJS divergence provide non-asymptotic upper bounds on the expected code length. The paper presents strictly positive lower bounds on EJS divergence, and hence non-asymptotic upper bounds on the expected code length, for the following two coding schemes: variable-length posterior matching and MaxEJS coding scheme which is based on a greedy maximization of the EJS divergence. As an asymptotic corollary of the main results, this paper also provides a rate-reliability test. Variable-length coding schemes that satisfy the condition(s) of the test for parameters $R$ and $E$, are guaranteed to achieve rate $R$ and error exponent $E$. The results are specialized for posterior matching and MaxEJS to obtain deterministic one-phase coding schemes achieving capacity and optimal error exponent. For the special case of symmetric binary-input channels, simpler deterministic schemes of optimal performance are proposed and analyzed.

cs.IT↗

arXiv2014

Searching with Measurement Dependent Noise

Consider a target moving with a constant velocity on a unit-circumference circle, starting from an arbitrary location. To acquire the target, any region of the circle can be probed for its presence, but the associated measurement noise increases with the size of the probed region. We are interested in the expected time required to find the target to within some given resolution and error probability. For a known velocity, we characterize the optimal tradeoff between time and resolution (i.e., maximal rate), and show that in contrast to the case of constant measurement noise, measurement dependent noise incurs a multiplicative gap between adaptive search and non-adaptive search. Moreover, our adaptive scheme attains the optimal rate-reliability tradeoff. We further show that for optimal non-adaptive search, accounting for an unknown velocity incurs a factor of two in rate.

cs.IT↗

arXiv2014

Distributed Learning of Distributions via Social Sampling

A protocol for distributed estimation of discrete distributions is proposed. Each agent begins with a single sample from the distribution, and the goal is to learn the empirical distribution of the samples. The protocol is based on a simple message-passing model motivated by communication in social networks. Agents sample a message randomly from their current estimates of the distribution, resulting in a protocol with quantized messages. Using tools from stochastic approximation, the algorithm is shown to converge almost surely. Examples illustrate three regimes with different consensus phenomena. Simulations demonstrate this convergence and give some insight into the effect of network topology.

math.OC↗

arXiv2013

Active sequential hypothesis testing

Consider a decision maker who is responsible to dynamically collect observations so as to enhance his information about an underlying phenomena of interest in a speedy manner while accounting for the penalty of wrong declaration. Due to the sequential nature of the problem, the decision maker relies on his current information state to adaptively select the most ``informative'' sensing action among the available ones. In this paper, using results in dynamic programming, lower bounds for the optimal total cost are established. The lower bounds characterize the fundamental limits on the maximum achievable information acquisition rate and the optimal reliability. Moreover, upper bounds are obtained via an analysis of two heuristic policies for dynamic selection of actions. It is shown that the first proposed heuristic achieves asymptotic optimality, where the notion of asymptotic optimality, due to Chernoff, implies that the relative difference between the total cost achieved by the proposed policy and the optimal total cost approaches zero as the penalty of wrong declaration (hence the number of collected samples) increases. The second heuristic is shown to achieve asymptotic optimality only in a limited setting such as the problem of a noisy dynamic search. However, by considering the dependency on the number of hypotheses, under a technical condition, this second heuristic is shown to achieve a nonzero information acquisition rate, establishing a lower bound for the maximum achievable rate and error exponent. In the case of a noisy dynamic search with size-independent noise, the obtained nonzero rate and error exponent are shown to be maximum.

cs.IT↗

arXiv2013

Noisy Bayesian Active Learning

We consider the problem of noisy Bayesian active learning, where we are given a finite set of functions $\mathcal{H}$, a sample space $\mathcal{X}$, and a label set $\mathcal{L}$. One of the functions in $\mathcal{H}$ assigns labels to samples in $\mathcal{X}$. The goal is to identify the function that generates the labels even though the result of a label query on a sample is corrupted by independent noise. More precisely, the objective is to declare one of the functions in $\mathcal{H}$ as the true label generating function with high confidence using as few label queries as possible, by selecting the queries adaptively and in a strategic manner. Previous work in Bayesian active learning considers Generalized Binary Search, and its variants for the noisy case, and analyzes the number of queries required by these sampling strategies. In this paper, we show that these schemes are, in general, suboptimal. Instead we propose and analyze an alternative strategy for sample collection. Our sampling strategy is motivated by a connection between Bayesian active learning and active hypothesis testing, and is based on querying the label of a sample which maximizes the Extrinsic Jensen-Shannon divergence at each step. We provide upper and lower bounds on the performance of this sampling strategy, and show that these bounds are better than previous bounds.

cs.IT↗

arXiv2012

Sequentiality and Adaptivity Gains in Active Hypothesis Testing

Consider a decision maker who is responsible to collect observations so as to enhance his information in a speedy manner about an underlying phenomena of interest. The policies under which the decision maker selects sensing actions can be categorized based on the following two factors: i) sequential vs. non-sequential; ii) adaptive vs. non-adaptive. Non-sequential policies collect a fixed number of observation samples and make the final decision afterwards; while under sequential policies, the sample size is not known initially and is determined by the observation outcomes. Under adaptive policies, the decision maker relies on the previous collected samples to select the next sensing action; while under non-adaptive policies, the actions are selected independent of the past observation outcomes. In this paper, performance bounds are provided for the policies in each category. Using these bounds, sequentiality gain and adaptivity gain, i.e., the gains of sequential and adaptive selection of actions are characterized.

cs.IT↗

arXiv2012

Opportunistic Routing with Congestion Diversity in Wireless Ad-hoc Networks

We consider the problem of routing packets across a multi-hop network consisting of multiple sources of traffic and wireless links while ensuring bounded expected delay. Each packet transmission can be overheard by a random subset of receiver nodes among which the next relay is selected opportunistically. The main challenge in the design of minimum-delay routing policies is balancing the trade-off between routing the packets along the shortest paths to the destination and distributing traffic according to the maximum backpressure. Combining important aspects of shortest path and backpressure routing, this paper provides a systematic development of a distributed opportunistic routing policy with congestion diversity ({D-ORCD}). {D-ORCD} uses a measure of draining time to opportunistically identify and route packets along the paths with an expected low overall congestion. {D-ORCD} is proved to ensure a bounded expected delay for all networks and under any admissible traffic. Furthermore, this paper proposes a practical implementation which empirically optimizes critical algorithm parameters and their effects on delay as well as protocol overhead. Realistic Qualnet simulations for 802.11-based networks demonstrate a significant improvement in the average delay over comparative solutions in the literature. %Finally, various practical modifications to {D-ORCD} are proposed and their performance are evaluated.

cs.NI↗

arXiv2011

Linear Sum Capacity for Gaussian Multiple Access Channels with Feedback

The capacity region of the N-sender Gaussian multiple access channel with feedback is not known in general. This paper studies the class of linear-feedback codes that includes (nonlinear) nonfeedback codes at one extreme and the linear-feedback codes by Schalkwijk and Kailath, Ozarow, and Kramer at the other extreme. The linear-feedback sum-capacity C_L(N,P) under symmetric power constraints P is characterized, the maximum sum-rate achieved by linear-feedback codes when each sender has the equal block power constraint P. In particular, it is shown that Kramer's code achieves this linear-feedback sum-capacity. The proof involves the dependence balance condition introduced by Hekstra and Willems and extended by Kramer and Gastpar, and the analysis of the resulting nonconvex optimization problem via a Lagrange dual formulation. Finally, an observation is presented based on the properties of the conditional maximal correlation---an extension of the Hirschfeld--Gebelein--Renyi maximal correlation---which reinforces the conjecture that Kramer's code achieves not only the linear-feedback sum-capacity, but also the sum-capacity itself (the maximum sum-rate achieved by arbitrary feedback codes).

cs.IT↗

arXiv2011

A General Class of Throughput Optimal Routing Policies in Multi-hop Wireless Networks

This paper considers the problem of throughput optimal routing/scheduling in a multi-hop constrained queueing network with random connectivity whose special case includes opportunistic multi-hop wireless networks and input-queued switch fabrics. The main challenge in the design of throughput optimal routing policies is closely related to identifying appropriate and universal Lyapunov functions with negative expected drift. The few well-known throughput optimal policies in the literature are constructed using simple quadratic or exponential Lyapunov functions of the queue backlogs and as such they seek to balance the queue backlogs across network independent of the topology. By considering a class of continuous, differentiable, and piece-wise quadratic Lyapunov functions, this paper provides a large class of throughput optimal routing policies. The proposed class of Lyapunov functions allow for the routing policy to control the traffic along short paths for a large portion of state-space while ensuring a negative expected drift. This structure enables the design of a large class of routing policies. In particular, and in addition to recovering the throughput optimality of the well known backpressure routing policy, an opportunistic routing policy with congestion diversity is proved to be throughput optimal.

math.OC↗

arXiv2009

Optimality of Myopic Sensing in Multi-Channel Opportunistic Access

We consider opportunistic communications over multiple channels where the state ("good" or "bad") of each channel evolves as independent and identically distributed Markov processes. A user, with limited sensing and access capability, chooses one channel to sense and subsequently access (based on the sensed channel state) in each time slot. A reward is obtained when the user senses and accesses a "good" channel. The objective is to design the optimal channel selection policy that maximizes the expected reward accrued over time. This problem can be generally cast as a Partially Observable Markov Decision Process (POMDP) or a restless multi-armed bandit process, to which optimal solutions are often intractable. We show in this paper that the myopic policy, with a simple and robust structure, achieves optimality under certain conditions. This result finds applications in opportunistic communications in fading environment, cognitive radio networks for spectrum overlay, and resource-constrained jamming and anti-jamming.

cs.NI↗

arXiv2009

Many-Sources Large Deviations for Max-Weight Scheduling

In this paper, a many-sources large deviations principle (LDP) for the transient workload of a multi-queue single-server system is established where the service rates are chosen from a compact, convex and coordinate-convex rate region and where the service discipline is the max-weight policy. Under the assumption that the arrival processes satisfy a many-sources LDP, this is accomplished by employing Garcia's extended contraction principle that is applicable to quasi-continuous mappings. For the simplex rate-region, an LDP for the stationary workload is also established under the additional requirements that the scheduling policy be work-conserving and that the arrival processes satisfy certain mixing conditions. The LDP results can be used to calculate asymptotic buffer overflow probabilities accounting for the multiplexing gain, when the arrival process is an average of \emph{i.i.d.} processes. The rate function for the stationary workload is expressed in term of the rate functions of the finite-horizon workloads when the arrival processes have \emph{i.i.d.} increments.

math.PR↗

arXiv2008

High-SNR Analysis of Outage-Limited Communications with Bursty and Delay-Limited Information

This work analyzes the high-SNR asymptotic error performance of outage-limited communications with fading, where the number of bits that arrive at the transmitter during any time slot is random but the delivery of bits at the receiver must adhere to a strict delay limitation. Specifically, bit errors are caused by erroneous decoding at the receiver or violation of the strict delay constraint. Under certain scaling of the statistics of the bit-arrival process with SNR, this paper shows that the optimal decay behavior of the asymptotic total probability of bit error depends on how fast the burstiness of the source scales down with SNR. If the source burstiness scales down too slowly, the total probability of error is asymptotically dominated by delay-violation events. On the other hand, if the source burstiness scales down too quickly, the total probability of error is asymptotically dominated by channel-error events. However, at the proper scaling, where the burstiness scales linearly with 1/sqrt(log SNR) and at the optimal coding duration and transmission rate, the occurrences of channel errors and delay-violation errors are asymptotically balanced. In this latter case, the optimal exponent of the total probability of error reveals a tradeoff that addresses the question of how much of the allowable time and rate should be used for gaining reliability over the channel and how much for accommodating the burstiness with delay constraints.

cs.IT↗