arXiv subjects

Bo Ji

Publications and source records attributed to Bo Ji.

At least 73 records · Page 4Linked to original sources

arXiv2022

On Kernelized Multi-Armed Bandits with Constraints

We study a stochastic bandit problem with a general unknown reward function and a general unknown constraint function. Both functions can be non-linear (even non-convex) and are assumed to lie in a reproducing kernel Hilbert space (RKHS) with a bounded norm. This kernelized bandit setup strictly generalizes standard multi-armed bandits and linear bandits. In contrast to safety-type hard constraints studied in prior works, we consider soft constraints that may be violated in any round as long as the cumulative violations are small, which is motivated by various practical applications. Our ultimate goal is to study how to utilize the nature of soft constraints to attain a finer complexity-regret-constraint trade-off in the kernelized bandit setting. To this end, leveraging primal-dual optimization, we propose a general framework for both algorithm design and performance analysis. This framework builds upon a novel sufficient condition, which not only is satisfied under general exploration strategies, including \emph{upper confidence bound} (UCB), \emph{Thompson sampling} (TS), and new ones based on \emph{random exploration}, but also enables a unified analysis for showing both sublinear regret and sublinear or even zero constraint violation. We demonstrate the superior performance of our proposed algorithms via numerical experiments based on both synthetic and real-world datasets. Along the way, we also make the first detailed comparison between two popular methods for analyzing constrained bandits and Markov decision processes (MDPs) by discussing the key difference and some subtleties in the analysis, which could be of independent interest to the communities.

cs.LG↗

arXiv2022

Placement and Allocation of Virtual Network Functions: Multi-dimensional Case

Network function virtualization (NFV) is an emerging design paradigm that replaces physical middlebox devices with software modules running on general purpose commodity servers. While gradually transitioning to NFV, Internet service providers face the problem of where to introduce NFV in order to make the most benefit of that; here, we measure the benefit by the amount of traffic that can be served in an NFV-enabled network. This problem is non-trivial as it is composed of two challenging subproblems: 1) placement of nodes to support virtual network functions (referred to as VNF-nodes); 2) allocation of the VNF-nodes' resources to network flows. This problem has been studied for the one-dimensional setting, where all network flows require one network function, which requires a unit of resource to process a unit of flow. In this work, we consider the multi-dimensional setting, where flows must be processed by multiple network functions, which require a different amount of each resource to process a unit of flow. The multi-dimensional setting introduces new challenges in addition to those of the one-dimensional setting (e.g., NP-hardness and non-submodularity) and also makes the resource allocation subproblem a multi-dimensional generalization of the generalized assignment problem with assignment restrictions. To address these difficulties, we propose a novel two-level relaxation method that allows us to draw a connection to the sequence submodular theory and utilize the property of sequence submodularity along with the primal-dual technique to design two approximation algorithms. We further prove that the proposed algorithms have a non-trivial approximation ratio that depends on the number of VNF-nodes, resources, and a measure of the available resource compared to flow demand. Finally, we perform trace-driven simulations to show the effectiveness of the proposed algorithms.

cs.NI↗

arXiv2022

GADGET: Online Resource Optimization for Scheduling Ring-All-Reduce Learning Jobs

Fueled by advances in distributed deep learning (DDL), recent years have witnessed a rapidly growing demand for resource-intensive distributed/parallel computing to process DDL computing jobs. To resolve network communication bottleneck and load balancing issues in distributed computing, the so-called ``ring-all-reduce'' decentralized architecture has been increasingly adopted to remove the need for dedicated parameter servers. To date, however, there remains a lack of theoretical understanding on how to design resource optimization algorithms for efficiently scheduling ring-all-reduce DDL jobs in computing clusters. This motivates us to fill this gap by proposing a series of new resource scheduling designs for ring-all-reduce DDL jobs. Our contributions in this paper are three-fold: i) We propose a new resource scheduling analytical model for ring-all-reduce deep learning, which covers a wide range of objectives in DDL performance optimization (e.g., excessive training avoidance, energy efficiency, fairness); ii) Based on the proposed performance analytical model, we develop an efficient resource scheduling algorithm called GADGET (greedy ring-all-reduce distributed graph embedding technique), which enjoys a provable strong performance guarantee; iii) We conduct extensive trace-driven experiments to demonstrate the effectiveness of the GADGET approach and its superiority over the state of the art.

cs.DC↗

arXiv2022

Joint Placement and Allocation of VNF Nodes with Budget and Capacity Constraints

With the advent of Network Function Virtualization (NFV), network services that traditionally run on proprietary dedicated hardware can now be realized using Virtual Network Functions (VNFs) that are hosted on general-purpose commodity hardware. This new network paradigm offers a great flexibility to Internet service providers (ISPs) for efficiently operating their networks (collecting network statistics, enforcing management policies, etc.). However, introducing NFV requires an investment to deploy VNFs at certain network nodes (called VNF-nodes), which has to account for practical constraints such as the deployment budget and the VNF-node capacity. To that end, it is important to design a joint VNF-nodes placement and capacity allocation algorithm that can maximize the total amount of network flows that are fully processed by the VNF-nodes while respecting such practical constraints. In contrast to most prior work that often neglects either the budget constraint or the capacity constraint, we explicitly consider both of them. We prove that accounting for these constraints introduces several new challenges. Specifically, we prove that the studied problem is not only NP-hard but also non-submodular. To address these challenges, we introduce a novel relaxation method such that the objective function of the relaxed placement subproblem becomes submodular. Leveraging this useful submodular property, we propose two algorithms that achieve an approximation ratio of $\frac{1}{2}(1-1/e)$ and $\frac{1}{3}(1-1/e)$ for the original non-relaxed problem, respectively. Finally, we corroborate the effectiveness of the proposed algorithms through extensive evaluations using trace-driven simulations.

cs.NI↗

arXiv2021

Only Train Once: A One-Shot Neural Network Training And Pruning Framework

Structured pruning is a commonly used technique in deploying deep neural networks (DNNs) onto resource-constrained devices. However, the existing pruning methods are usually heuristic, task-specified, and require an extra fine-tuning procedure. To overcome these limitations, we propose a framework that compresses DNNs into slimmer architectures with competitive performances and significant FLOPs reductions by Only-Train-Once (OTO). OTO contains two keys: (i) we partition the parameters of DNNs into zero-invariant groups, enabling us to prune zero groups without affecting the output; and (ii) to promote zero groups, we then formulate a structured-sparsity optimization problem and propose a novel optimization algorithm, Half-Space Stochastic Projected Gradient (HSPG), to solve it, which outperforms the standard proximal methods on group sparsity exploration and maintains comparable convergence. To demonstrate the effectiveness of OTO, we train and compress full models simultaneously from scratch without fine-tuning for inference speedup and parameter reduction, and achieve state-of-the-art results on VGG16 for CIFAR10, ResNet50 for CIFAR10 and Bert for SQuAD and competitive result on ResNet50 for ImageNet. The source code is available at https://github.com/tianyic/only_train_once.

cs.LG↗

arXiv2021

Federated Learning with Fair Worker Selection: A Multi-Round Submodular Maximization Approach

In this paper, we study the problem of fair worker selection in Federated Learning systems, where fairness serves as an incentive mechanism that encourages more workers to participate in the federation. Considering the achieved training accuracy of the global model as the utility of the selected workers, which is typically a monotone submodular function, we formulate the worker selection problem as a new multi-round monotone submodular maximization problem with cardinality and fairness constraints. The objective is to maximize the time-average utility over multiple rounds subject to an additional fairness requirement that each worker must be selected for a certain fraction of time. While the traditional submodular maximization with a cardinality constraint is already a well-known NP-Hard problem, the fairness constraint in the multi-round setting adds an extra layer of difficulty. To address this novel challenge, we propose three algorithms: Fair Continuous Greedy (FairCG1 and FairCG2) and Fair Discrete Greedy (FairDG), all of which satisfy the fairness requirement whenever feasible. Moreover, we prove nontrivial lower bounds on the achieved time-average utility under FairCG1 and FairCG2. In addition, by giving a higher priority to fairness, FairDG ensures a stronger short-term fairness guarantee, which holds in every round. Finally, we perform extensive simulations to verify the effectiveness of the proposed algorithms in terms of the time-average utility and fairness satisfaction.

cs.GT↗

arXiv2021

A Sum-of-Ratios Multi-Dimensional-Knapsack Decomposition for DNN Resource Scheduling

In recent years, to sustain the resource-intensive computational needs for training deep neural networks (DNNs), it is widely accepted that exploiting the parallelism in large-scale computing clusters is critical for the efficient deployments of DNN training jobs. However, existing resource schedulers for traditional computing clusters are not well suited for DNN training, which results in unsatisfactory job completion time performance. The limitations of these resource scheduling schemes motivate us to propose a new computing cluster resource scheduling framework that is able to leverage the special layered structure of DNN jobs and significantly improve their job completion times. Our contributions in this paper are three-fold: i) We develop a new resource scheduling analytical model by considering DNN's layered structure, which enables us to analytically formulate the resource scheduling optimization problem for DNN training in computing clusters; ii) Based on the proposed performance analytical model, we then develop an efficient resource scheduling algorithm based on the widely adopted parameter-server architecture using a sum-of-ratios multi-dimensional-knapsack decomposition (SMD) method to offer strong performance guarantee; iii) We conduct extensive numerical experiments to demonstrate the effectiveness of the proposed schedule algorithm and its superior performance over the state of the art.

cs.DC↗

arXiv2021

Half-Space Proximal Stochastic Gradient Method for Group-Sparsity Regularized Problem

Optimizing with group sparsity is significant in enhancing model interpretability in machining learning applications, e.g., feature selection, compressed sensing and model compression. However, for large-scale stochastic training problems, effective group sparsity exploration are typically hard to achieve. Particularly, the state-of-the-art stochastic optimization algorithms usually generate merely dense solutions. To overcome this shortage, we propose a stochastic method -- Half-space Stochastic Projected Gradient (HSPG) method to search solutions of high group sparsity while maintain the convergence. Initialized by a simple Prox-SG Step, the HSPG method relies on a novel Half-Space Step to substantially boost the sparsity level. Numerically, HSPG demonstrates its superiority in deep neural networks, e.g., VGG16, ResNet18 and MobileNetV1, by computing solutions of higher group sparsity, competitive objective values and generalization accuracy.

math.OC↗

arXiv2020

Neural Network Compression Via Sparse Optimization

The compression of deep neural networks (DNNs) to reduce inference cost becomes increasingly important to meet realistic deployment requirements of various applications. There have been a significant amount of work regarding network compression, while most of them are heuristic rule-based or typically not friendly to be incorporated into varying scenarios. On the other hand, sparse optimization yielding sparse solutions naturally fits the compression requirement, but due to the limited study of sparse optimization in stochastic learning, its extension and application onto model compression is rarely well explored. In this work, we propose a model compression framework based on the recent progress on sparse stochastic optimization. Compared to existing model compression techniques, our method is effective and requires fewer extra engineering efforts to incorporate with varying applications, and has been numerically demonstrated on benchmark compression tasks. Particularly, we achieve up to 7.2 and 2.9 times FLOPs reduction with the same level of evaluation accuracy on VGG16 for CIFAR10 and ResNet50 for ImageNet compared to the baseline heavy models, respectively.

cs.LG↗

arXiv2020

Robust Sequence Submodular Maximization

Submodularity is an important property of set functions and has been extensively studied in the literature. It models set functions that exhibit a diminishing returns property, where the marginal value of adding an element to a set decreases as the set expands. This notion has been generalized to considering sequence functions, where the order of adding elements plays a crucial role and determines the function value; the generalized notion is called sequence (or string) submodularity. In this paper, we study a new problem of robust sequence submodular maximization with cardinality constraints. The robustness is against the removal of a subset of elements in the selected sequence (e.g., due to malfunctions or adversarial attacks). Compared to robust submodular maximization for set function, new challenges arise when sequence functions are concerned. Specifically, there are multiple definitions of submodularity for sequence functions, which exhibit subtle yet critical differences. Another challenge comes from two directions of monotonicity: forward monotonicity and backward monotonicity, both of which are important to proving performance guarantees. To address these unique challenges, we design two robust greedy algorithms: while one algorithm achieves a constant approximation ratio but is robust only against the removal of a subset of contiguous elements, the other is robust against the removal of an arbitrary subset of the selected elements but requires a stronger assumption and achieves an approximation ratio that depends on the number of the removed elements. Finally, we generalize the analyses to considering sequence functions under weaker assumptions based on approximate versions of sequence submodularity and backward monotonicity

cs.DS↗

arXiv2020

Anti-Aging Scheduling in Single-Server Queues: A Systematic and Comparative Study

The Age-of-Information (AoI) is a new performance metric recently proposed for measuring the freshness of information in information-update systems. In this work, we conduct a systematic and comparative study to investigate the impact of scheduling policies on the AoI performance in single-server queues and provide useful guidelines for the design of AoI-efficient scheduling policies. Specifically, we first perform extensive simulations to demonstrate that the update-size information can be leveraged for achieving a substantially improved AoI compared to non-size-based (or arrival-time-based) policies. Then, by utilizing both the update-size and arrival-time information, we propose three AoI-based policies. Observing improved AoI performance of policies that allow service preemption and that prioritize informative updates, we further propose preemptive, informative, AoI-based scheduling policies. Our simulation results show that such policies empirically achieve the best AoI performance among all the considered policies. However, compared to the best delay-efficient policies (such as Shortest-Remaining-Processing-Time (SRPT)), the AoI improvement is rather marginal in the settings with exogenous arrivals. Interestingly, we also prove sample-path equivalence between some size-based policies and AoI-based policies. This provides an intuitive explanation for why some size-based policies (such as SRPT) achieve a very good AoI performance.

cs.IT↗

arXiv2020

Orthant Based Proximal Stochastic Gradient Method for $\ell_1$-Regularized Optimization

Sparsity-inducing regularization problems are ubiquitous in machine learning applications, ranging from feature selection to model compression. In this paper, we present a novel stochastic method -- Orthant Based Proximal Stochastic Gradient Method (OBProx-SG) -- to solve perhaps the most popular instance, i.e., the l1-regularized problem. The OBProx-SG method contains two steps: (i) a proximal stochastic gradient step to predict a support cover of the solution; and (ii) an orthant step to aggressively enhance the sparsity level via orthant face projection. Compared to the state-of-the-art methods, e.g., Prox-SG, RDA and Prox-SVRG, the OBProx-SG not only converges to the global optimal solutions (in convex scenario) or the stationary points (in non-convex scenario), but also promotes the sparsity of the solutions substantially. Particularly, on a large number of convex problems, OBProx-SG outperforms the existing methods comprehensively in the aspect of sparsity exploration and objective values. Moreover, the experiments on non-convex deep neural networks, e.g., MobileNetV1 and ResNet18, further demonstrate its superiority by achieving the solutions of much higher sparsity without sacrificing generalization accuracy.

math.OC↗

arXiv2020

Generative Adversarial Network for Handwritten Text

Generative adversarial networks (GANs) have proven hugely successful in variety of applications of image processing. However, generative adversarial networks for handwriting is relatively rare somehow because of difficulty of handling sequential handwriting data by Convolutional Neural Network (CNN). In this paper, we propose a handwriting generative adversarial network framework (HWGANs) for synthesizing handwritten stroke data. The main features of the new framework include: (i) A discriminator consists of an integrated CNN-Long-Short-Term- Memory (LSTM) based feature extraction with Path Signature Features (PSF) as input and a Feedforward Neural Network (FNN) based binary classifier; (ii) A recurrent latent variable model as generator for synthesizing sequential handwritten data. The numerical experiments show the effectivity of the new model. Moreover, comparing with sole handwriting generator, the HWGANs synthesize more natural and realistic handwritten text.

cs.LG↗

arXiv2019

Combinatorial Sleeping Bandits with Fairness Constraints

The multi-armed bandit (MAB) model has been widely adopted for studying many practical optimization problems (network resource allocation, ad placement, crowdsourcing, etc.) with unknown parameters. The goal of the player here is to maximize the cumulative reward in the face of uncertainty. However, the basic MAB model neglects several important factors of the system in many real-world applications, where multiple arms can be simultaneously played and an arm could sometimes be "sleeping". Besides, ensuring fairness is also a key design concern in practice. To that end, we propose a new Combinatorial Sleeping MAB model with Fairness constraints, called CSMAB-F, aiming to address the aforementioned crucial modeling issues. The objective is now to maximize the reward while satisfying the fairness requirement of a minimum selection fraction for each individual arm. To tackle this new problem, we extend an online learning algorithm, UCB, to deal with a critical tradeoff between exploitation and exploration and employ the virtual queue technique to properly handle the fairness constraints. By carefully integrating these two techniques, we develop a new algorithm, called Learning with Fairness Guarantee (LFG), for the CSMAB-F problem. Further, we rigorously prove that not only LFG is feasibility-optimal, but it also has a time-average regret upper bounded by $\frac{N}{2η}+\frac{β_1\sqrt{mNT\log{T}}+β_2 N}{T}$, where N is the total number of arms, m is the maximum number of arms that can be simultaneously played, T is the time horizon, $β_1$ and $β_2$ are constants, and $η$ is a design parameter that we can tune. Finally, we perform extensive simulations to corroborate the effectiveness of the proposed algorithm. Interestingly, the simulation results reveal an important tradeoff between the regret and the speed of convergence to a point satisfying the fairness constraints.

cs.LG↗

arXiv2019

Towards the Tradeoff Between Service Performance and Information Freshness

The last decade has witnessed an unprecedented growth in the demand for data-driven real-time services. These services are fueled by emerging applications that require rapidly injecting data streams and computing updated analytics results in real-time. In many of such applications, the computing resources are often shared for processing both updates from information sources and queries from end users. This requires joint scheduling of updates and queries because the service provider needs to make a critical decision upon receiving a user query: either it responds immediately with currently available but possibly stale information, or it first processes new updates and then responds with fresher information. Hence, the tradeoff between service performance and information freshness naturally arises in this context. To that end, we propose a simple single-server two-queue model that captures the coupled scheduling of updates and queries and aim to design scheduling policies that can properly address the important tradeoff between performance and freshness. Specifically, we consider the response time as a performance metric and the Age of Information (AoI) as a freshness metric. After demonstrating the limitations of the simplest FCFS policy, we propose two threshold-based policies: the Query-k policy that prioritizes queries and the Update-k policy that prioritizes updates. Then, we rigorously analyze both the response time and the Peak AoI (PAoI) of the threshold-based policies. Further, we propose the Joint-(M,N) policy, which allows flexibly prioritizing updates or queries through choosing different values of two thresholds M and N. Finally, we conduct simulations to evaluate the response time and the PAoI of the proposed policies. The results show that our proposed threshold-based policies can effectively control the balance between performance and freshness.

cs.PF↗

arXiv2018

Shortest Path and Maximum Flow Problems Under Service Function Chaining Constraints

With the advent of Network Function Virtualization (NFV), Physical Network Functions (PNFs) are gradually being replaced by Virtual Network Functions (VNFs) that are hosted on general purpose servers. Depending on the call flows for specific services, the packets need to pass through an ordered set of network functions (physical or virtual) called Service Function Chains (SFC) before reaching the destination. Conceivably for the next few years during this transition, these networks would have a mix of PNFs and VNFs, which brings an interesting mix of network problems that are studied in this paper: (1) How to find an SFC-constrained shortest path between any pair of nodes? (2) What is the achievable SFC-constrained maximum flow? (3) How to place the VNFs such that the cost (the number of nodes to be virtualized) is minimized, while the maximum flow of the original network can still be achieved even under the SFC constraint? In this work, we will try to address such emerging questions. First, for the SFC-constrained shortest path problem, we propose a transformation of the network graph to minimize the computational complexity of subsequent applications of any shortest path algorithm. Second, we formulate the SFC-constrained maximum flow problem as a fractional multicommodity flow problem, and develop a combinatorial algorithm for a special case of practical interest. Third, we prove that the VNFs placement problem is NP-hard and present an alternative Integer Linear Programming (ILP) formulation. Finally, we conduct simulations to elucidate our theoretical results.

cs.NI↗

arXiv2017

Node-based Service-Balanced Scheduling for Provably Guaranteed Throughput and Evacuation Time Performance

This paper focuses on the design of provably efficient online link scheduling algorithms for multi-hop wireless networks. We consider single-hop traffic and the one-hop interference model. The objective is twofold: 1) maximizing the throughput when the flow sources continuously inject packets into the network, and 2) minimizing the evacuation time when there are no future packet arrivals. The prior work mostly employs the link-based approach, which leads to throughput-efficient algorithms but often does not guarantee satisfactory evacuation time performance. In this paper, we propose a novel Node-based Service-Balanced (NSB) online scheduling algorithm. NSB aims to give scheduling opportunities to heavily congested nodes in a balanced manner, by maximizing the total weight of the scheduled nodes in each scheduling cycle, where the weight of a node is determined by its workload and whether the node was scheduled in the previous scheduling cycle(s). We rigorously prove that NSB guarantees to achieve an efficiency ratio no worse (or no smaller) than 2/3 for the throughput and an approximation ratio no worse (or no greater) than 3/2 for the evacuation time. It is remarkable that NSB is both throughput-optimal and evacuation-time-optimal if the underlying network graph is bipartite. Further, we develop a lower-complexity NSB algorithm, called LC-NSB, which provides the same performance guarantees as NSB. Finally, we conduct numerical experiments to elucidate our theoretical results.

cs.NI↗

arXiv2017

The Power of Waiting for More than One Response in Minimizing the Age-of-Information

The Age-of-Information (AoI) has recently been proposed as an important metric for investigating the timeliness performance in information-update systems. Prior studies on AoI optimization often consider a Push model, which is concerned about when and how to "push" (i.e., generate and transmit) the updated information to the user. In stark contrast, in this paper we introduce a new Pull model, which is more relevant for certain applications (such as the real-time stock quotes service), where a user sends requests to the servers to proactively "pull" the information of interest. Moreover, we propose to employ request replication to reduce the AoI. Interestingly, we find that under this new Pull model, replication schemes capture a novel tradeoff between different levels of information freshness and different response times across the servers, which can be exploited to minimize the expected AoI at the user's side. Specifically, assuming Poisson updating process at the servers and exponentially distributed response time, we derive a closedform formula for computing the expected AoI and obtain the optimal number of responses to wait for to minimize the expected AoI. Finally, we conduct numerical simulations to elucidate our theoretical results. Our findings show that waiting for more than one response can significantly reduce the AoI in most scenarios.

cs.NI↗