arXiv subjects

Ion Stoica

Publications and source records attributed to Ion Stoica.

At least 235 records · Page 13Linked to original sources

arXiv2020

IMPACT: Importance Weighted Asynchronous Architectures with Clipped Target Networks

The practical usage of reinforcement learning agents is often bottlenecked by the duration of training time. To accelerate training, practitioners often turn to distributed reinforcement learning architectures to parallelize and accelerate the training process. However, modern methods for scalable reinforcement learning (RL) often tradeoff between the throughput of samples that an RL agent can learn from (sample throughput) and the quality of learning from each sample (sample efficiency). In these scalable RL architectures, as one increases sample throughput (i.e. increasing parallelization in IMPALA), sample efficiency drops significantly. To address this, we propose a new distributed reinforcement learning algorithm, IMPACT. IMPACT extends IMPALA with three changes: a target network for stabilizing the surrogate objective, a circular buffer, and truncated importance sampling. In discrete action-space environments, we show that IMPACT attains higher reward and, simultaneously, achieves up to 30% decrease in training wall-time than that of IMPALA. For continuous control environments, IMPACT trains faster than existing scalable agents while preserving the sample efficiency of synchronous PPO.

cs.LG↗

arXiv2020

HyperSched: Dynamic Resource Reallocation for Model Development on a Deadline

Prior research in resource scheduling for machine learning training workloads has largely focused on minimizing job completion times. Commonly, these model training workloads collectively search over a large number of parameter values that control the learning process in a hyperparameter search. It is preferable to identify and maximally provision the best-performing hyperparameter configuration (trial) to achieve the highest accuracy result as soon as possible. To optimally trade-off evaluating multiple configurations and training the most promising ones by a fixed deadline, we design and build HyperSched -- a dynamic application-level resource scheduler to track, identify, and preferentially allocate resources to the best performing trials to maximize accuracy by the deadline. HyperSched leverages three properties of a hyperparameter search workload over-looked in prior work - trial disposability, progressively identifiable rankings among different configurations, and space-time constraints - to outperform standard hyperparameter search algorithms across a variety of benchmarks.

cs.DC↗

arXiv2020

NeuroVectorizer: End-to-End Vectorization with Deep Reinforcement Learning

One of the key challenges arising when compilers vectorize loops for today's SIMD-compatible architectures is to decide if vectorization or interleaving is beneficial. Then, the compiler has to determine how many instructions to pack together and how many loop iterations to interleave. Compilers are designed today to use fixed-cost models that are based on heuristics to make vectorization decisions on loops. However, these models are unable to capture the data dependency, the computation graph, or the organization of instructions. Alternatively, software engineers often hand-write the vectorization factors of every loop. This, however, places a huge burden on them, since it requires prior experience and significantly increases the development time. In this work, we explore a novel approach for handling loop vectorization and propose an end-to-end solution using deep reinforcement learning (RL). We conjecture that deep RL can capture different instructions, dependencies, and data structures to enable learning a sophisticated model that can better predict the actual performance cost and determine the optimal vectorization factors. We develop an end-to-end framework, from code to vectorization, that integrates deep RL in the LLVM compiler. Our proposed framework takes benchmark codes as input and extracts the loop codes. These loop codes are then fed to a loop embedding generator that learns an embedding for these loops. Finally, the learned embeddings are used as input to a Deep RL agent, which determines the vectorization factors for all the loops. We further extend our framework to support multiple supervised learning methods. We evaluate our approaches against the currently used LLVM vectorizer and loop polyhedral optimization techniques. Our experiments show 1.29X-4.73X performance speedup compared to baseline and only 3% worse than the brute-force search on a wide range of benchmarks.

cs.DC↗

arXiv2019

Hierarchical Variational Imitation Learning of Control Programs

Autonomous agents can learn by imitating teacher demonstrations of the intended behavior. Hierarchical control policies are ubiquitously useful for such learning, having the potential to break down structured tasks into simpler sub-tasks, thereby improving data efficiency and generalization. In this paper, we propose a variational inference method for imitation learning of a control policy represented by parametrized hierarchical procedures (PHP), a program-like structure in which procedures can invoke sub-procedures to perform sub-tasks. Our method discovers the hierarchical structure in a dataset of observation-action traces of teacher demonstrations, by learning an approximate posterior distribution over the latent sequence of procedure calls and terminations. Samples from this learned distribution then guide the training of the hierarchical control policy. We identify and demonstrate a novel benefit of variational inference in the context of hierarchical imitation learning: in decomposing the policy into simpler procedures, inference can leverage acausal information that is unused by other methods. Training PHP with variational inference outperforms LSTM baselines in terms of data efficiency and generalization, requiring less than half as much data to achieve a 24% error rate in executing the bubble sort algorithm, and to achieve no error in executing Karel programs.

cs.LG↗

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

Deep Unsupervised Cardinality Estimation

Cardinality estimation has long been grounded in statistical tools for density estimation. To capture the rich multivariate distributions of relational tables, we propose the use of a new type of high-capacity statistical model: deep autoregressive models. However, direct application of these models leads to a limited estimator that is prohibitively expensive to evaluate for range or wildcard predicates. To produce a truly usable estimator, we develop a Monte Carlo integration scheme on top of autoregressive models that can efficiently handle range queries with dozens of dimensions or more. Like classical synopses, our estimator summarizes the data without supervision. Unlike previous solutions, we approximate the joint data distribution without any independence assumptions. Evaluated on real-world datasets and compared against real systems and dominant families of techniques, our estimator achieves single-digit multiplicative error at tail, an up to 90$\times$ accuracy improvement over the second best method, and is space- and runtime-efficient.

cs.DB↗

arXiv2019

Blink: Fast and Generic Collectives for Distributed ML

Model parameter synchronization across GPUs introduces high overheads for data-parallel training at scale. Existing parameter synchronization protocols cannot effectively leverage available network resources in the face of ever increasing hardware heterogeneity. To address this, we propose Blink, a collective communication library that dynamically generates optimal communication primitives by packing spanning trees. We propose techniques to minimize the number of trees generated and extend Blink to leverage heterogeneous communication channels for faster data transfers. Evaluations show that compared to the state-of-the-art (NCCL), Blink can achieve up to 8x faster model synchronization, and reduce end-to-end training time for image classification tasks by up to 40%.

cs.DC↗

arXiv2019

A View on Deep Reinforcement Learning in System Optimization

Many real-world systems problems require reasoning about the long term consequences of actions taken to configure and manage the system. These problems with delayed and often sequentially aggregated reward, are often inherently reinforcement learning problems and present the opportunity to leverage the recent substantial advances in deep reinforcement learning. However, in some cases, it is not clear why deep reinforcement learning is a good fit for the problem. Sometimes, it does not perform better than the state-of-the-art solutions. And in other cases, random search or greedy algorithms could outperform deep reinforcement learning. In this paper, we review, discuss, and evaluate the recent trends of using deep reinforcement learning in system optimization. We propose a set of essential metrics to guide future works in evaluating the efficacy of using deep reinforcement learning in system optimization. Our evaluation includes challenges, the types of problems, their formulation in the deep reinforcement learning setting, embedding, the model used, efficiency, and robustness. We conclude with a discussion on open challenges and potential directions for pushing further the integration of reinforcement learning in system optimization.

cs.LG↗

arXiv2019

Helen: Maliciously Secure Coopetitive Learning for Linear Models

Many organizations wish to collaboratively train machine learning models on their combined datasets for a common benefit (e.g., better medical research, or fraud detection). However, they often cannot share their plaintext datasets due to privacy concerns and/or business competition. In this paper, we design and build Helen, a system that allows multiple parties to train a linear model without revealing their data, a setting we call coopetitive learning. Compared to prior secure training systems, Helen protects against a much stronger adversary who is malicious and can compromise m-1 out of m parties. Our evaluation shows that Helen can achieve up to five orders of magnitude of performance improvement when compared to training using an existing state-of-the-art secure multi-party computation framework.

cs.CR↗

arXiv2019

Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules

A key challenge in leveraging data augmentation for neural network training is choosing an effective augmentation policy from a large search space of candidate operations. Properly chosen augmentation policies can lead to significant generalization improvements; however, state-of-the-art approaches such as AutoAugment are computationally infeasible to run for the ordinary user. In this paper, we introduce a new data augmentation algorithm, Population Based Augmentation (PBA), which generates nonstationary augmentation policy schedules instead of a fixed augmentation policy. We show that PBA can match the performance of AutoAugment on CIFAR-10, CIFAR-100, and SVHN, with three orders of magnitude less overall compute. On CIFAR-10 we achieve a mean test error of 1.46%, which is a slight improvement upon the current state-of-the-art. The code for PBA is open source and is available at https://github.com/arcelien/pba.

cs.CV↗

arXiv2019

Harmonia: Near-Linear Scalability for Replicated Storage with In-Network Conflict Detection

Distributed storage employs replication to mask failures and improve availability. However, these systems typically exhibit a hard tradeoff between consistency and performance. Ensuring consistency introduces coordination overhead, and as a result the system throughput does not scale with the number of replicas. We present Harmonia, a replicated storage architecture that exploits the capability of new-generation programmable switches to obviate this tradeoff by providing near-linear scalability without sacrificing consistency. To achieve this goal, Harmonia detects read-write conflicts in the network, which enables any replica to serve reads for objects with no pending writes. Harmonia implements this functionality at line rate, thus imposing no performance overhead. We have implemented a prototype of Harmonia on a cluster of commodity servers connected by a Barefoot Tofino switch, and have integrated it with Redis. We demonstrate the generality of our approach by supporting a variety of replication protocols, including primary-backup, chain replication, Viewstamped Replication, and NOPaxos. Experimental results show that Harmonia improves the throughput of these protocols by up to 10X for a replication factor of 10, providing near-linear scalability up to the limit of our testbed.

cs.DC↗

arXiv2019

AutoPhase: Compiler Phase-Ordering for High Level Synthesis with Deep Reinforcement Learning

The performance of the code generated by a compiler depends on the order in which the optimization passes are applied. In high-level synthesis, the quality of the generated circuit relates directly to the code generated by the front-end compiler. Choosing a good order--often referred to as the phase-ordering problem--is an NP-hard problem. In this paper, we evaluate a new technique to address the phase-ordering problem: deep reinforcement learning. We implement a framework in the context of the LLVM compiler to optimize the ordering for HLS programs and compare the performance of deep reinforcement learning to state-of-the-art algorithms that address the phase-ordering problem. Overall, our framework runs one to two orders of magnitude faster than these algorithms, and achieves a 16% improvement in circuit performance over the -O3 compiler flag.

cs.PL↗

arXiv2019

Neural Packet Classification

Packet classification is a fundamental problem in computer networking. This problem exposes a hard tradeoff between the computation and state complexity, which makes it particularly challenging. To navigate this tradeoff, existing solutions rely on complex hand-tuned heuristics, which are brittle and hard to optimize. In this paper, we propose a deep reinforcement learning (RL) approach to solve the packet classification problem. There are several characteristics that make this problem a good fit for Deep RL. First, many of the existing solutions are iteratively building a decision tree by splitting nodes in the tree. Second, the effects of these actions (e.g., splitting nodes) can only be evaluated once we are done with building the tree. These two characteristics are naturally captured by the ability of RL to take actions that have sparse and delayed rewards. Third, it is computationally efficient to generate data traces and evaluate decision trees, which alleviate the notoriously high sample complexity problem of Deep RL algorithms. Our solution, NeuroCuts, uses succinct representations to encode state and action space, and efficiently explore candidate decision trees to optimize for a global objective. It produces compact decision trees optimized for a specific set of rules and a given performance metric, such as classification time, memory footprint, or a combination of the two. Evaluation on ClassBench shows that NeuroCuts outperforms existing hand-crafted algorithms in classification time by 18% at the median, and reduces both time and memory footprint by up to 3x.

cs.NI↗

arXiv2019

DistCache: Provable Load Balancing for Large-Scale Storage Systems with Distributed Caching

Load balancing is critical for distributed storage to meet strict service-level objectives (SLOs). It has been shown that a fast cache can guarantee load balancing for a clustered storage system. However, when the system scales out to multiple clusters, the fast cache itself would become the bottleneck. Traditional mechanisms like cache partition and cache replication either result in load imbalance between cache nodes or have high overhead for cache coherence. We present DistCache, a new distributed caching mechanism that provides provable load balancing for large-scale storage systems. DistCache co-designs cache allocation with cache topology and query routing. The key idea is to partition the hot objects with independent hash functions between cache nodes in different layers, and to adaptively route queries with the power-of-two-choices. We prove that DistCache enables the cache throughput to increase linearly with the number of cache nodes, by unifying techniques from expander graphs, network flows, and queuing theory. DistCache is a general solution that can be applied to many storage systems. We demonstrate the benefits of DistCache by providing the design, implementation, and evaluation of the use case for emerging switch-based caching.

cs.DC↗

arXiv2019

Cloud Programming Simplified: A Berkeley View on Serverless Computing

Serverless cloud computing handles virtually all the system administration operations needed to make it easier for programmers to use the cloud. It provides an interface that greatly simplifies cloud programming, and represents an evolution that parallels the transition from assembly language to high-level programming languages. This paper gives a quick history of cloud computing, including an accounting of the predictions of the 2009 Berkeley View of Cloud Computing paper, explains the motivation for serverless computing, describes applications that stretch the current limits of serverless, and then lists obstacles and research opportunities required for serverless computing to fulfill its full potential. Just as the 2009 paper identified challenges for the cloud and predicted they would be addressed and that cloud use would accelerate, we predict these issues are solvable and that serverless computing will grow to dominate the future of cloud computing.

cs.OS↗

arXiv2019

The OoO VLIW JIT Compiler for GPU Inference

Current trends in Machine Learning~(ML) inference on hardware accelerated devices (e.g., GPUs, TPUs) point to alarmingly low utilization. As ML inference is increasingly time-bounded by tight latency SLOs, increasing data parallelism is not an option. The need for better efficiency motivates GPU multiplexing. Furthermore, existing GPU programming abstractions force programmers to micro-manage GPU resources in an early-binding, context-free fashion. We propose a VLIW-inspired Out-of-Order (OoO) Just-in-Time (JIT) compiler that coalesces and reorders execution kernels at runtime for throughput-optimal device utilization while satisfying latency SLOs. We quantify the inefficiencies of space-only and time-only multiplexing alternatives and demonstrate an achievable 7.7x opportunity gap through spatial coalescing.

cs.DC↗

arXiv2019

Learning to Optimize Join Queries With Deep Reinforcement Learning

Exhaustive enumeration of all possible join orders is often avoided, and most optimizers leverage heuristics to prune the search space. The design and implementation of heuristics are well-understood when the cost model is roughly linear, and we find that these heuristics can be significantly suboptimal when there are non-linearities in cost. Ideally, instead of a fixed heuristic, we would want a strategy to guide the search space in a more data-driven way---tailoring the search to a specific dataset and query workload. Recognizing the link between classical Dynamic Programming enumeration methods and recent results in Reinforcement Learning (RL), we propose a new method for learning optimized join search strategies. We present our RL-based DQ optimizer, which currently optimizes select-project-join blocks. We implement three versions of DQ to illustrate the ease of integration into existing DBMSes: (1) A version built on top of Apache Calcite, (2) a version integrated into PostgreSQL, and (3) a version integrated into SparkSQL. Our extensive evaluation shows that DQ achieves plans with optimization costs and query execution times competitive with the native query optimizer in each system, but can execute significantly faster after learning (often by orders of magnitude).

cs.DB↗

arXiv2018

Dynamic Space-Time Scheduling for GPU Inference

Serving deep neural networks in latency critical interactive settings often requires GPU acceleration. However, the small batch sizes typical in online inference results in poor GPU utilization, a potential performance gap which GPU resource sharing can address. In this paper, we explore several techniques to leverage both temporal and spatial multiplexing to improve GPU utilization for deep learning inference workloads. We evaluate the performance trade-offs of each approach with respect to resource-efficiency, latency predictability, and isolation when compared with conventional batched inference. Our experimental analysis suggests up to a 5x potential for improved utilization through the exploration of more advanced spatial and temporal multiplexing strategies. Our preliminary prototype of a dynamic space-time scheduler demonstrates a 3.23x floating-point throughput increase over space-only multiplexing and a 7.73x increase over time-only multiplexing for convolutions, while also providing better isolation and latency predictability.

cs.DC↗