arXiv subjects

Dan Alistarh

Publications and source records attributed to Dan Alistarh.

At least 145 records · Page 8Linked to original sources

arXiv2021

Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks

The growing energy and performance costs of deep learning have driven the community to reduce the size of neural networks by selectively pruning components. Similarly to their biological counterparts, sparse networks generalize just as well, if not better than, the original dense networks. Sparsity can reduce the memory footprint of regular networks to fit mobile devices, as well as shorten training time for ever growing networks. In this paper, we survey prior work on sparsity in deep learning and provide an extensive tutorial of sparsification for both inference and training. We describe approaches to remove and add elements of neural networks, different training strategies to achieve model sparsity, and mechanisms to exploit sparsity in practice. Our work distills ideas from more than 300 research papers and provides guidance to practitioners who wish to utilize sparsity today, as well as to researchers whose goal is to push the frontier forward. We include the necessary background on mathematical methods in sparsification, describe phenomena such as early structure adaptation, the intricate relations between sparsity and the training process, and show techniques for achieving acceleration on real hardware. We also define a metric of pruned parameter efficiency that could serve as a baseline for comparison of different sparse networks. We close by speculating on how sparsity can improve future workloads and outline major open problems in the field.

cs.LG↗

arXiv2021

Relaxed Scheduling for Scalable Belief Propagation

The ability to leverage large-scale hardware parallelism has been one of the key enablers of the accelerated recent progress in machine learning. Consequently, there has been considerable effort invested into developing efficient parallel variants of classic machine learning algorithms. However, despite the wealth of knowledge on parallelization, some classic machine learning algorithms often prove hard to parallelize efficiently while maintaining convergence. In this paper, we focus on efficient parallel algorithms for the key machine learning task of inference on graphical models, in particular on the fundamental belief propagation algorithm. We address the challenge of efficiently parallelizing this classic paradigm by showing how to leverage scalable relaxed schedulers in this context. We present an extensive empirical study, showing that our approach outperforms previous parallel belief propagation implementations both in terms of scalability and in terms of wall-clock convergence time, on a range of practical applications.

cs.DC↗

arXiv2020

Compressive Sensing Using Iterative Hard Thresholding with Low Precision Data Representation: Theory and Applications

Modern scientific instruments produce vast amounts of data, which can overwhelm the processing ability of computer systems. Lossy compression of data is an intriguing solution, but comes with its own drawbacks, such as potential signal loss, and the need for careful optimization of the compression ratio. In this work, we focus on a setting where this problem is especially acute: compressive sensing frameworks for interferometry and medical imaging. We ask the following question: can the precision of the data representation be lowered for all inputs, with recovery guarantees and practical performance? Our first contribution is a theoretical analysis of the normalized Iterative Hard Thresholding (IHT) algorithm when all input data, meaning both the measurement matrix and the observation vector are quantized aggressively. We present a variant of low precision normalized {IHT} that, under mild conditions, can still provide recovery guarantees. The second contribution is the application of our quantization framework to radio astronomy and magnetic resonance imaging. We show that lowering the precision of the data can significantly accelerate image recovery. We evaluate our approach on telescope data and samples of brain images using CPU and FPGA implementations achieving up to a 9x speed-up with negligible loss of recovery quality.

stat.ML↗

arXiv2020

WoodFisher: Efficient Second-Order Approximation for Neural Network Compression

Second-order information, in the form of Hessian- or Inverse-Hessian-vector products, is a fundamental tool for solving optimization problems. Recently, there has been significant interest in utilizing this information in the context of deep neural networks; however, relatively little is known about the quality of existing approximations in this context. Our work examines this question, identifies issues with existing approaches, and proposes a method called WoodFisher to compute a faithful and efficient estimate of the inverse Hessian. Our main application is to neural network compression, where we build on the classic Optimal Brain Damage/Surgeon framework. We demonstrate that WoodFisher significantly outperforms popular state-of-the-art methods for one-shot pruning. Further, even when iterative, gradual pruning is considered, our method results in a gain in test accuracy over the state-of-the-art approaches, for pruning popular neural networks (like ResNet-50, MobileNetV1) trained on standard image classification datasets such as ImageNet ILSVRC. We examine how our method can be extended to take into account first-order information, as well as illustrate its ability to automatically set layer-wise pruning thresholds and perform compression in the limited-data regime. The code is available at the following link, https://github.com/IST-DASLab/WoodFisher.

cs.LG↗

arXiv2020

Adaptive Gradient Quantization for Data-Parallel SGD

Many communication-efficient variants of SGD use gradient quantization schemes. These schemes are often heuristic and fixed over the course of training. We empirically observe that the statistics of gradients of deep models change during the training. Motivated by this observation, we introduce two adaptive quantization schemes, ALQ and AMQ. In both schemes, processors update their compression schemes in parallel by efficiently computing sufficient statistics of a parametric distribution. We improve the validation accuracy by almost 2% on CIFAR-10 and 1% on ImageNet in challenging low-cost communication setups. Our adaptive methods are also significantly more robust to the choice of hyperparameters.

cs.LG↗

arXiv2020

The Splay-List: A Distribution-Adaptive Concurrent Skip-List

The design and implementation of efficient concurrent data structures have seen significant attention. However, most of this work has focused on concurrent data structures providing good \emph{worst-case} guarantees. In real workloads, objects are often accessed at different rates, since access distributions may be non-uniform. Efficient distribution-adaptive data structures are known in the sequential case, e.g. the splay-trees; however, they often are hard to translate efficiently in the concurrent case. In this paper, we investigate distribution-adaptive concurrent data structures and propose a new design called the splay-list. At a high level, the splay-list is similar to a standard skip-list, with the key distinction that the height of each element adapts dynamically to its access rate: popular elements ``move up,'' whereas rarely-accessed elements decrease in height. We show that the splay-list provides order-optimal amortized complexity bounds for a subset of operations while being amenable to efficient concurrent implementation. Experimental results show that the splay-list can leverage distribution-adaptivity to improve on the performance of classic concurrent designs, and can outperform the only previously-known distribution-adaptive design in certain settings.

cs.DC↗

arXiv2020

Why Extension-Based Proofs Fail

We introduce extension-based proofs, a class of impossibility proofs that includes valency arguments. They are modelled as an interaction between a prover and a protocol. Using proofs based on combinatorial topology, it has been shown that it is impossible to deterministically solve k-set agreement among n > k > 1 processes in a wait-free manner in certain asynchronous models. However, it was unknown whether proofs based on simpler techniques were possible. We show that this impossibility result cannot be obtained for one of these models by an extension-based proof and, hence, extension-based proofs are limited in power.

cs.DC↗

arXiv2020

Asynchronous Optimization Methods for Efficient Training of Deep Neural Networks with Guarantees

Asynchronous distributed algorithms are a popular way to reduce synchronization costs in large-scale optimization, and in particular for neural network training. However, for nonsmooth and nonconvex objectives, few convergence guarantees exist beyond cases where closed-form proximal operator solutions are available. As most popular contemporary deep neural networks lead to nonsmooth and nonconvex objectives, there is now a pressing need for such convergence guarantees. In this paper, we analyze for the first time the convergence of stochastic asynchronous optimization for this general class of objectives. In particular, we focus on stochastic subgradient methods allowing for block variable partitioning, where the shared-memory-based model is asynchronously updated by concurrent processes. To this end, we first introduce a probabilistic model which captures key features of real asynchronous scheduling between concurrent processes; under this model, we establish convergence with probability one to an invariant set for stochastic subgradient methods with momentum. From the practical perspective, one issue with the family of methods we consider is that it is not efficiently supported by machine learning frameworks, as they mostly focus on distributed data-parallel strategies. To address this, we propose a new implementation strategy for shared-memory based training of deep neural networks, whereby concurrent parameter servers are utilized to train a partitioned but shared model in single- and multi-GPU settings. Based on this implementation, we achieve on average 1.2x speed-up in comparison to state-of-the-art training methods for popular image classification tasks without compromising accuracy.

math.OC↗

arXiv2020

On the Sample Complexity of Adversarial Multi-Source PAC Learning

We study the problem of learning from multiple untrusted data sources, a scenario of increasing practical relevance given the recent emergence of crowdsourcing and collaborative learning paradigms. Specifically, we analyze the situation in which a learning system obtains datasets from multiple sources, some of which might be biased or even adversarially perturbed. It is known that in the single-source case, an adversary with the power to corrupt a fixed fraction of the training data can prevent PAC-learnability, that is, even in the limit of infinitely much training data, no learning system can approach the optimal test error. In this work we show that, surprisingly, the same is not true in the multi-source setting, where the adversary can arbitrarily corrupt a fixed fraction of the data sources. Our main results are a generalization bound that provides finite-sample guarantees for this learning setting, as well as corresponding lower bounds. Besides establishing PAC-learnability our results also show that in a cooperative learning setting sharing data with other parties has provable benefits, even if some participants are malicious.

cs.LG↗

arXiv2020

Elastic Consistency: A General Consistency Model for Distributed Stochastic Gradient Descent

Machine learning has made tremendous progress in recent years, with models matching or even surpassing humans on a series of specialized tasks. One key element behind the progress of machine learning in recent years has been the ability to train machine learning models in large-scale distributed shared-memory and message-passing environments. Many of these models are trained employing variants of stochastic gradient descent (SGD) based optimization. In this paper, we introduce a general consistency condition covering communication-reduced and asynchronous distributed SGD implementations. Our framework, called elastic consistency enables us to derive convergence bounds for a variety of distributed SGD methods used in practice to train large-scale machine learning models. The proposed framework de-clutters the implementation-specific convergence analysis and provides an abstraction to derive convergence bounds. We utilize the framework to analyze a sparsification scheme for distributed SGD methods in an asynchronous setting for convex and non-convex objectives. We implement the distributed SGD variant to train deep CNN models in an asynchronous shared-memory setting. Empirical results show that error-feedback may not necessarily help in improving the convergence of sparsified asynchronous distributed SGD, which corroborates an insight suggested by our convergence analysis.

cs.LG↗

arXiv2020

Fast General Distributed Transactions with Opacity using Global Time

Transactions can simplify distributed applications by hiding data distribution, concurrency, and failures from the application developer. Ideally the developer would see the abstraction of a single large machine that runs transactions sequentially and never fails. This requires the transactional subsystem to provide opacity (strict serializability for both committed and aborted transactions), as well as transparent fault tolerance with high availability. As even the best abstractions are unlikely to be used if they perform poorly, the system must also provide high performance. Existing distributed transactional designs either weaken this abstraction or are not designed for the best performance within a data center. This paper extends the design of FaRM - which provides strict serializability only for committed transactions - to provide opacity while maintaining FaRM's high throughput, low latency, and high availability within a modern data center. It uses timestamp ordering based on real time with clocks synchronized to within tens of microseconds across a cluster, and a failover protocol to ensure correctness across clock master failures. FaRM with opacity can commit 5.4 million neworder transactions per second when running the TPC-C transaction mix on 90 machines with 3-way replication.

cs.DC↗

arXiv2020

Stochastic Gradient Langevin with Delayed Gradients

Stochastic Gradient Langevin Dynamics (SGLD) ensures strong guarantees with regards to convergence in measure for sampling log-concave posterior distributions by adding noise to stochastic gradient iterates. Given the size of many practical problems, parallelizing across several asynchronously running processors is a popular strategy for reducing the end-to-end computation time of stochastic optimization algorithms. In this paper, we are the first to investigate the effect of asynchronous computation, in particular, the evaluation of stochastic Langevin gradients at delayed iterates, on the convergence in measure. For this, we exploit recent results modeling Langevin dynamics as solving a convex optimization problem on the space of measures. We show that the rate of convergence in measure is not significantly affected by the error caused by the delayed gradient information used for computation, suggesting significant potential for speedup in wall clock time. We confirm our theoretical results with numerical experiments on some practical problems.

cs.LG↗

arXiv2020

Efficiency Guarantees for Parallel Incremental Algorithms under Relaxed Schedulers

Several classic problems in graph processing and computational geometry are solved via incremental algorithms, which split computation into a series of small tasks acting on shared state, which gets updated progressively. While the sequential variant of such algorithms usually specifies a fixed (but sometimes random) order in which the tasks should be performed, a standard approach to parallelizing such algorithms is to relax this constraint to allow for out-of-order parallel execution. This is the case for parallel implementations of Dijkstra's single-source shortest-paths algorithm (SSSP), and for parallel Delaunay mesh triangulation. While many software frameworks parallelize incremental computation in this way, it is still not well understood whether this relaxed ordering approach can still provide any complexity guarantees. In this paper, we address this problem, and analyze the efficiency guarantees provided by a range of incremental algorithms when parallelized via relaxed schedulers. We show that, for algorithms such as Delaunay mesh triangulation and sorting by insertion, schedulers with a maximum relaxation factor of $k$ in terms of the maximum priority inversion allowed will introduce a maximum amount of wasted work of $O(log(n) poly (k) ), $ where $n$ is the number of tasks to be executed. For SSSP, we show that the additional work is $O(poly (k) d_{max} / w_{min}), $ where $d_{\max}$ is the maximum distance between two nodes, and $w_{min}$ is the minimum such distance. In practical settings where $n \gg k$, this suggests that the overheads of relaxation will be outweighed by the improved scalability of the relaxed scheduler. On the negative side, we provide lower bounds showing that certain algorithms will inherently incur a non-trivial amount of wasted work due to scheduler relaxation, even for relatively benign relaxed schedulers.

cs.DS↗

arXiv2020

Dynamic Averaging Load Balancing on Cycles

We consider the following dynamic load-balancing process: given an underlying graph $G$ with $n$ nodes, in each step $t\geq 0$, one unit of load is created, and placed at a randomly chosen graph node. In the same step, the chosen node picks a random neighbor, and the two nodes balance their loads by averaging them. We are interested in the expected gap between the minimum and maximum loads at nodes as the process progresses, and its dependence on $n$ and on the graph structure. Similar variants of the above graphical balanced allocation process have been studied by Peres, Talwar, and Wieder, and by Sauerwald and Sun for regular graphs. These authors left as open the question of characterizing the gap in the case of \emph{cycle graphs} in the \emph{dynamic} case, where weights are created during the algorithm's execution. For this case, the only known upper bound is of $\mathcal{O}( n \log n )$, following from a majorization argument due to Peres, Talwar, and Wieder, which analyzes a related graphical allocation process. In this paper, we provide an upper bound of $\mathcal{O} ( \sqrt n \log n )$ on the expected gap of the above process for cycles of length $n$. We introduce a new potential analysis technique, which enables us to bound the difference in load between $k$-hop neighbors on the cycle, for any $k \leq n / 2$. We complement this with a "gap covering" argument, which bounds the maximum value of the gap by bounding its value across all possible subsets of a certain structure, and recursively bounding the gaps within each subset. We provide analytical and experimental evidence that our upper bound on the gap is tight up to a logarithmic factor.

cs.DS↗

arXiv2020

Powerset Convolutional Neural Networks

We present a novel class of convolutional neural networks (CNNs) for set functions, i.e., data indexed with the powerset of a finite set. The convolutions are derived as linear, shift-equivariant functions for various notions of shifts on set functions. The framework is fundamentally different from graph convolutions based on the Laplacian, as it provides not one but several basic shifts, one for each element in the ground set. Prototypical experiments with several set function classification tasks on synthetic datasets and on datasets derived from real-world hypergraphs demonstrate the potential of our new powerset CNNs.

cs.LG↗

arXiv2020

Analysis and Evaluation of Non-Blocking Interpolation Search Trees

We start by summarizing the recently proposed implementation of the first non-blocking concurrent interpolation search tree (C-IST) data structure. We then analyze the individual operations of the C-IST, and show that they are correct and linearizable. We furthermore show that lookup (and several other non-destructive operations) are wait-free, and that the insert and delete operations are lock-free. We continue by showing that the C-IST has the following properties. For arbitrary key distributions, this data structure ensures worst-case $O(\log n + p)$ amortized time for search, insertion and deletion traversals. When the input key distributions are smooth, lookups run in expected $O(\log \log n + p)$ time, and insertion and deletion run in expected amortized $O(\log \log n + p)$ time, where $p$ is a bound on the number of threads. Finally, we present an extended experimental evaluation of the non-blocking IST performance.

cs.DS↗

arXiv2019

MLSys: The New Frontier of Machine Learning Systems

Machine learning (ML) techniques are enjoying rapidly increasing adoption. However, designing and implementing the systems that support ML models in real-world deployments remains a significant obstacle, in large part due to the radically different development and deployment profile of modern ML methods, and the range of practical concerns that come with broader adoption. We propose to foster a new systems machine learning research community at the intersection of the traditional systems and ML communities, focused on topics such as hardware systems for ML, software systems for ML, and ML optimized for metrics beyond predictive accuracy. To do this, we describe a new conference, MLSys, that explicitly targets research at the intersection of systems and machine learning with a program committee split evenly between experts in systems and ML, and an explicit focus on topics at the intersection of the two.

cs.LG↗

arXiv2019

In Search of the Fastest Concurrent Union-Find Algorithm

Union-Find (or Disjoint-Set Union) is one of the fundamental problems in computer science; it has been well-studied from both theoretical and practical perspectives in the sequential case. Recently, there has been mounting interest in analyzing this problem in the concurrent scenario, and several asymptotically-efficient algorithms have been proposed. Yet, to date, there is very little known about the practical performance of concurrent Union-Find. This work addresses this gap. We evaluate and analyze the performance of several concurrent Union-Find algorithms and optimization strategies across a wide range of platforms (Intel, AMD, and ARM) and workloads (social, random, and road networks, as well as integrations into more complex algorithms). We first observe that, due to the limited computational cost, the number of induced cache misses is the critical determining factor for the performance of existing algorithms. We introduce new techniques to reduce this cost by storing node priorities implicitly and by using plain reads and writes in a way that does not affect the correctness of the algorithms. Finally, we show that Union-Find implementations are an interesting application for Transactional Memory (TM): one of the fastest algorithm variants we discovered is a sequential one that uses coarse-grained locking with the lock elision optimization to reduce synchronization cost and increase scalability.

cs.DS↗