NORD (Neural Operations Research & Development) is an open source distributed deep learning architectural research framework, based on PyTorch, MPI and Horovod. It aims to make research of deep architectures easier for experts of different domains, in order to accelerate the process of finding better architectures, as well as study the best architectures generated for different datasets. Although currently under heavy development, the framework aims to allow the easy implementation of different design and optimization method families (optimization algorithms, meta-heuristics, reinforcement learning etc.) as well as the fair comparison between them. Furthermore, due to the computational resources required in order to optimize and evaluate network architectures, it leverage the use of distributed computing, while aiming to minimize the researcher's overhead required to implement it. Moreover, it strives to make the creation of architectures more intuitive, by implementing network descriptors, allowing to separately define the architecture's nodes and connections. In this paper, we present the framework's current state of development, while presenting its basic concepts, providing simple examples as well as their experimental results.
In this paper we present the FolksoDriven Cloud (FDC) built on Cloud and on Semantic technologies. Cloud computing has emerged in these recent years as the new paradigm for the provision of on-demand distributed computing resources. Semantic Web can be used for relationship between different data and descriptions of services to annotate provenance of repositories on ontologies. The FDC service is composed of a back-end which submits and monitors the documents, and a user front-end which allows users to schedule on-demand operations and to watch the progress of running processes. The impact of the proposed method is illustrated on a user since its inception.
A recent take towards Federated Analytics (FA), which allows analytical insights of distributed datasets, reuses the Federated Learning (FL) infrastructure to evaluate the summary of model performances across the training devices. However, the current realization of FL adopts single server-multiple client architecture with limited scope for FA, which often results in learning models with poor generalization, i.e., an ability to handle new/unseen data, for real-world applications. Moreover, a hierarchical FL structure with distributed computing platforms demonstrates incoherent model performances at different aggregation levels. Therefore, we need to design a robust learning mechanism than the FL that (i) unleashes a viable infrastructure for FA and (ii) trains learning models with better generalization capability. In this work, we adopt the novel democratized learning (Dem-AI) principles and designs to meet these objectives. Firstly, we show the hierarchical learning structure of the proposed edge-assisted democratized learning mechanism, namely Edge-DemLearn, as a practical framework to empower generalization capability in support of FA. Secondly, we validate Edge-DemLearn as a flexible model training mechanism to build a distributed control and aggregation methodology in regions by leveraging the distributed computing infrastructure. The distributed edge computing servers construct regional models, minimize the communication loads, and ensure distributed data analytic application's scalability. To that end, we adhere to a near-optimal two-sided many-to-one matching approach to handle the combinatorial constraints in Edge-DemLearn and solve it for fast knowledge acquisition with optimization of resource allocation and associations between multiple servers and devices. Extensive simulation results on real datasets demonstrate the effectiveness of the proposed methods.
Unmanned aerial vehicles (UAVs) are capable of serving as flying base stations (BSs) for supporting data collection, artificial intelligence (AI) model training, and wireless communications. However, due to the privacy concerns of devices and limited computation or communication resource of UAVs, it is impractical to send raw data of devices to UAV servers for model training. Moreover, due to the dynamic channel condition and heterogeneous computing capacity of devices in UAV-enabled networks, the reliability and efficiency of data sharing require to be further improved. In this paper, we develop an asynchronous federated learning (AFL) framework for multi-UAV-enabled networks, which can provide asynchronous distributed computing by enabling model training locally without transmitting raw sensitive data to UAV servers. The device selection strategy is also introduced into the AFL framework to keep the low-quality devices from affecting the learning efficiency and accuracy. Moreover, we propose an asynchronous advantage actor-critic (A3C) based joint device selection, UAVs placement, and resource management algorithm to enhance the federated convergence speed and accuracy. Simulation results demonstrate that our proposed framework and algorithm achieve higher learning accuracy and faster federated execution time compared to other existing solutions.
The issue of potential privacy leakage during centralized AI's model training has drawn intensive concern from the public. A Parallel and Distributed Computing (or PDC) scheme, termed Federated Learning (FL), has emerged as a new paradigm to cope with the privacy issue by allowing clients to perform model training locally, without the necessity to upload their personal sensitive data. In FL, the number of clients could be sufficiently large, but the bandwidth available for model distribution and re-upload is quite limited, making it sensible to only involve part of the volunteers to participate in the training process. The client selection policy is critical to an FL process in terms of training efficiency, the final model's quality as well as fairness. In this paper, we will model the fairness guaranteed client selection as a Lyapunov optimization problem and then a C2MAB-based method is proposed for estimation of the model exchange time between each client and the server, based on which we design a fairness guaranteed algorithm termed RBCS-F for problem-solving. The regret of RBCS-F is strictly bounded by a finite constant, justifying its theoretical feasibility. Barring the theoretical results, more empirical data can be derived from our real training experiments on public datasets.
We study exploration in stochastic multi-armed bandits when we have access to a divisible resource that can be allocated in varying amounts to arm pulls. We focus in particular on the allocation of distributed computing resources, where we may obtain results faster by allocating more resources per pull, but might have reduced throughput due to nonlinear scaling. For example, in simulation-based scientific studies, an expensive simulation can be sped up by running it on multiple cores. This speed-up however, is partly offset by the communication among cores, which results in lower throughput than if fewer cores were allocated per trial to run more trials in parallel. In this paper, we explore these trade-offs in two settings. First, in a fixed confidence setting, we need to find the best arm with a given target success probability as quickly as possible. We propose an algorithm which trades off between information accumulation and throughput and show that the time taken can be upper bounded by the solution of a dynamic program whose inputs are the gaps between the sub-optimal and optimal arms. We also prove a matching hardness result. Second, we present an algorithm for a fixed deadline setting, where we are given a time deadline and need to maximize the probability of finding the best arm. We corroborate our theoretical insights with simulation experiments that show that the algorithms consistently match or outperform baseline algorithms on a variety of problem instances.
In this work, a heterogeneous set of wireless devices sharing a common access point collaborates to perform a set of tasks. Using the Map-Reduce distributed computing framework, the tasks are optimally distributed amongst the nodes with the objective of minimizing the total energy consumption of the nodes while satisfying a latency constraint. The derived optimal collaborative-computing scheme takes into account both the computing capabilities of the nodes and the strength of their communication links. Numerical simulations illustrate the benefits of the proposed optimal collaborative-computing scheme over a blind collaborative-computing scheme and the non-collaborative scenario, both in term of energy savings and achievable latency. The proposed optimal scheme also exhibits the interesting feature of allowing to trade energy for latency, and vice versa.
Codistillation has been proposed as a mechanism to share knowledge among concurrently trained models by encouraging them to represent the same function through an auxiliary loss. This contrasts with the more commonly used fully-synchronous data-parallel stochastic gradient descent methods, where different model replicas average their gradients (or parameters) at every iteration and thus maintain identical parameters. We investigate codistillation in a distributed training setup, complementing previous work which focused on extremely large batch sizes. Surprisingly, we find that even at moderate batch sizes, models trained with codistillation can perform as well as models trained with synchronous data-parallel methods, despite using a much weaker synchronization mechanism. These findings hold across a range of batch sizes and learning rate schedules, as well as different kinds of models and datasets. Obtaining this level of accuracy, however, requires properly accounting for the regularization effect of codistillation, which we highlight through several empirical observations. Overall, this work contributes to a better understanding of codistillation and how to best take advantage of it in a distributed computing environment.
We develop a modeling and simulation framework capable of massively parallel simulation of multicellular systems with spatially resolved stochastic kinetics in individual cells. By the use of operator-splitting we decouple the simulation of reaction-diffusion kinetics inside the cells from the simulation of molecular cell-cell interactions occurring on the boundaries between cells. This decoupling leverages the inherent scale separation in the underlying model to enable highly horizontally scalable parallel simulation, suitable for simulation on heterogeneous, distributed computing infrastructures such as public and private clouds. Thanks to its modular structure, our frameworks makes it possible to couple just any existing single-cell simulation software together with any cell signaling simulator. We exemplify the flexibility and scalability of the framework by using the popular single-cell simulation software eGFRD to construct and simulate a multicellular model of Notch-Delta signaling over OpenStack cloud infrastructure provided by the SNIC Science Cloud.
In living systems, we often see the emergence of the ingredients necessary for computation -- the capacity for information transmission, storage, and modification -- begging the question of how we may exploit or imitate such biological systems in unconventional computing applications. What can we gain from artificial life in the advancement of computing technology? Artificial life provides us with powerful tools for understanding the dynamic behavior of biological systems and capturing this behavior in manmade substrates. With this approach, we can move towards a new computing paradigm concerned with harnessing emergent computation in physical substrates not governed by the constraints of Moore's law and ultimately realize massively parallel and distributed computing technology. In this paper, we argue that the lens of artificial life offers valuable perspectives for the advancement of high-performance computing technology. We first present a brief foundational background on artificial life and some relevant tools that may be applicable to unconventional computing. Two specific substrates are then discussed in detail: biological neurons and ensembles of nanomagnets. These substrates are the focus of the authors' ongoing work, and they are illustrative of the two sides of the approach outlined here -- the close study of living systems and the construction of artificial systems to produce life-like behaviors. We conclude with a philosophical discussion on what we can learn from approaching computation with the curiosity inherent to the study of artificial life. The main contribution of this paper is to present the great potential of using artificial life methodologies to uncover and harness the inherent computational power of physical substrates toward applications in unconventional high-performance computing.
Solving partial differential equations (PDE) is an indispensable part of many branches of science as many processes can be modelled in terms of PDEs. However, recent numerical solvers require manual discretization of the underlying equation as well as sophisticated, tailored code for distributed computing. Scanning the parameters of the underlying model significantly increases the runtime as the simulations have to be cold-started for each parameter configuration. Machine Learning based surrogate models denote promising ways for learning complex relationship among input, parameter and solution. However, recent generative neural networks require lots of training data, i.e. full simulation runs making them costly. In contrast, we examine the applicability of continuous, mesh-free neural solvers for partial differential equations, physics-informed neural networks (PINNs) solely requiring initial/boundary values and validation points for training but no simulation data. The induced curse of dimensionality is approached by learning a domain decomposition that steers the number of neurons per unit volume and significantly improves runtime. Distributed training on large-scale cluster systems also promises great utilization of large quantities of GPUs which we assess by a comprehensive evaluation study. Finally, we discuss the accuracy of GatedPINN with respect to analytical solutions -- as well as state-of-the-art numerical solvers, such as spectral solvers.
Distributed computing is a standard way to scale up machine learning and data science algorithms to process large amounts of data. In such settings, avoiding communication amongst machines is paramount for achieving high performance. Rather than distribute the computation of existing algorithms, a common practice for avoiding communication is to compute local solutions or parameter estimates on each machine and then combine the results; in many convex optimization problems, even simple averaging of local solutions can work well. However, these schemes do not work when the local solutions are not unique. Spectral methods are a collection of such problems, where solutions are orthonormal bases of the leading invariant subspace of an associated data matrix, which are only unique up to rotation and reflections. Here, we develop a communication-efficient distributed algorithm for computing the leading invariant subspace of a data matrix. Our algorithm uses a novel alignment scheme that minimizes the Procrustean distance between local solutions and a reference solution, and only requires a single round of communication. For the important case of principal component analysis (PCA), we show that our algorithm achieves a similar error rate to that of a centralized estimator. We present numerical experiments demonstrating the efficacy of our proposed algorithm for distributed PCA, as well as other problems where solutions exhibit rotational symmetry, such as node embeddings for graph data and spectral initialization for quadratic sensing.
A distributed computing scenario is considered, where the computational power of a set of worker nodes is used to perform a certain computation task over a dataset that is dispersed among the workers. Lagrange coded computing (LCC), proposed by Yu et al., leverages the well-known Lagrange polynomial to perform polynomial evaluation of the dataset in such a scenario in an efficient parallel fashion while keeping the privacy of data amidst possible collusion of workers. This solution relies on quantizing the data into a finite field, so that Shamir's secret sharing, as one of its main building blocks, can be employed. Such a solution, however, is not properly scalable with the size of dataset, mainly due to computation overflows. To address such a critical issue, we propose a novel extension of LCC to the analog domain, referred to as analog LCC (ALCC). All the operations in the proposed ALCC protocol are done over the infinite fields of R/C but for practical implementations floating-point numbers are used. We characterize the privacy of data in ALCC, against any subset of colluding workers up to a certain size, in terms of the distinguishing security (DS) and the mutual information security (MIS) metrics. Also, the accuracy of outcome is characterized in a practical setting assuming operations are performed using floating-point numbers. Consequently, a fundamental trade-off between the accuracy of the outcome of ALCC and its privacy level is observed and is numerically evaluated. Moreover, we implement the proposed scheme to perform matrix-matrix multiplication over a batch of matrices. It is observed that ALCC is superior compared to the state-of-the-art LCC, implemented using fixed-point numbers, assuming both schemes use an equal number of bits to represent data symbols.
Recent progress in scientific machine learning (SciML) has opened up the possibility of training novel neural network architectures that solve complex partial differential equations (PDEs). Several (nearly data free) approaches have been recently reported that successfully solve PDEs, with examples including deep feed forward networks, generative networks, and deep encoder-decoder networks. However, practical adoption of these approaches is limited by the difficulty in training these models, especially to make predictions at large output resolutions ($\geq 1024 \times 1024$). Here we report on a software framework for data parallel distributed deep learning that resolves the twin challenges of training these large SciML models - training in reasonable time as well as distributing the storage requirements. Our framework provides several out of the box functionality including (a) loss integrity independent of number of processes, (b) synchronized batch normalization, and (c) distributed higher-order optimization methods. We show excellent scalability of this framework on both cloud as well as HPC clusters, and report on the interplay between bandwidth, network topology and bare metal vs cloud. We deploy this approach to train generative models of sizes hitherto not possible, showing that neural PDE solvers can be viably trained for practical applications. We also demonstrate that distributed higher-order optimization methods are $2-3\times$ faster than stochastic gradient-based methods and provide minimal convergence drift with higher batch-size.
Federated learning (FL) is a rapidly growing research field in machine learning. However, existing FL libraries cannot adequately support diverse algorithmic development; inconsistent dataset and model usage make fair algorithm comparison challenging. In this work, we introduce FedML, an open research library and benchmark to facilitate FL algorithm development and fair performance comparison. FedML supports three computing paradigms: on-device training for edge devices, distributed computing, and single-machine simulation. FedML also promotes diverse algorithmic research with flexible and generic API design and comprehensive reference baseline implementations (optimizer, models, and datasets). We hope FedML could provide an efficient and reproducible means for developing and evaluating FL algorithms that would benefit the FL research community. We maintain the source code, documents, and user community at https://fedml.ai.
The use of Artificial Intelligence (AI) to detect phishing emails is primarily dependent on large-scale centralized datasets, which opens it up to a myriad of privacy, trust, and legal issues. Moreover, organizations are loathed to share emails, given the risk of leakage of commercially sensitive information. So, it is uncommon to obtain sufficient emails to train a global AI model efficiently. Accordingly, privacy-preserving distributed and collaborative machine learning, particularly Federated Learning (FL), is a desideratum. Already prevalent in the healthcare sector, questions remain regarding the effectiveness and efficacy of FL-based phishing detection within the context of multi-organization collaborations. To the best of our knowledge, the work herein is the first to investigate the use of FL in email anti-phishing. This paper builds upon a deep neural network model, particularly RNN and BERT for phishing email detection. It analyzes the FL-entangled learning performance under various settings, including balanced and asymmetrical data distribution. Our results corroborate comparable performance statistics of FL in phishing email detection to centralized learning for balanced datasets, and low organization counts. Moreover, we observe a variation in performance when increasing organizational counts. For a fixed total email dataset, the global RNN based model suffers by a 1.8% accuracy drop when increasing organizational counts from 2 to 10. In contrast, BERT accuracy rises by 0.6% when going from 2 to 5 organizations. However, if we allow increasing the overall email dataset with the introduction of new organizations in the FL framework, the organizational level performance is improved by achieving a faster convergence speed. Besides, FL suffers in its overall global model performance due to highly unstable outputs if the email dataset distribution is highly asymmetric.
In recent years, substantial progress has been made on Graph Convolutional Networks (GCNs). However, the computing of GCN usually requires a large memory space for keeping the entire graph. In consequence, GCN is not flexible enough, especially for large scale graphs in complex real-world applications. Fortunately, methods based on Matrix Factorization (MF) naturally support constructing mini-batches, and thus are more friendly to distributed computing compared with GCN. Accordingly, in this paper, we analyze the connections between GCN and MF, and simplify GCN as matrix factorization with unitization and co-training. Furthermore, under the guidance of our analysis, we propose an alternative model to GCN named Unitized and Co-training Matrix Factorization (UCMF). Extensive experiments have been conducted on several real-world datasets. On the task of semi-supervised node classification, the experimental results illustrate that UCMF achieves similar or superior performances compared with GCN. Meanwhile, distributed UCMF significantly outperforms distributed GCN methods, which shows that UCMF can greatly benefit large scale and complex real-world applications. Moreover, we have also conducted experiments on a typical task of graph embedding, i.e., community detection, and the proposed UCMF model outperforms several representative graph embedding models.
Traditional methods for black box optimization require a considerable number of evaluations which can be time consuming, unpractical, and often unfeasible for many engineering applications that rely on accurate representations and expensive models to evaluate. Bayesian Optimization (BO) methods search for the global optimum by progressively (actively) learning a surrogate model of the objective function along the search path. Bayesian optimization can be accelerated through multifidelity approaches which leverage multiple black-box approximations of the objective functions that can be computationally cheaper to evaluate, but still provide relevant information to the search task. Further computational benefits are offered by the availability of parallel and distributed computing architectures whose optimal usage is an open opportunity within the context of active learning. This paper introduces the Resource Aware Active Learning (RAAL) strategy, a multifidelity Bayesian scheme to accelerate the optimization of black box functions. At each optimization step, the RAAL procedure computes the set of best sample locations and the associated fidelity sources that maximize the information gain to acquire during the parallel/distributed evaluation of the objective function, while accounting for the limited computational budget. The scheme is demonstrated for a variety of benchmark problems and results are discussed for both single fidelity and multifidelity settings. In particular we observe that the RAAL strategy optimally seeds multiple points at each iteration allowing for a major speed up of the optimization task.
Recent works have shown that agents facing independent instances of a stochastic $K$-armed bandit can collaborate to decrease regret. However, these works assume that each agent always recommends their individual best-arm estimates to other agents, which is unrealistic in envisioned applications (machine faults in distributed computing or spam in social recommendation systems). Hence, we generalize the setting to include $n$ honest and $m$ malicious agents who recommend best-arm estimates and arbitrary arms, respectively. We first show that even with a single malicious agent, existing collaboration-based algorithms fail to improve regret guarantees over a single-agent baseline. We propose a scheme where honest agents learn who is malicious and dynamically reduce communication with (i.e., "block") them. We show that collaboration indeed decreases regret for this algorithm, assuming $m$ is small compared to $K$ but without assumptions on malicious agents' behavior, thus ensuring that our algorithm is robust against any malicious recommendation strategy.
Trade-offs between accuracy and efficiency pervade law, public health, and other non-computing domains, which have developed policies to guide how to balance the two in conditions of uncertainty. While computer science also commonly studies accuracy-efficiency trade-offs, their policy implications remain poorly examined. Drawing on risk assessment practices in the US, we argue that, since examining these trade-offs has been useful for guiding governance in other domains, we need to similarly reckon with these trade-offs in governing computer systems. We focus our analysis on distributed machine learning systems. Understanding the policy implications in this area is particularly urgent because such systems, which include autonomous vehicles, tend to be high-stakes and safety-critical. We 1) describe how the trade-off takes shape for these systems, 2) highlight gaps between existing US risk assessment standards and what these systems require to be properly assessed, and 3) make specific calls to action to facilitate accountability when hypothetical risks concerning the accuracy-efficiency trade-off become realized as accidents in the real world. We close by discussing how such accountability mechanisms encourage more just, transparent governance aligned with public values.