arXiv subjects

Eiko Yoneki

Publications and source records attributed to Eiko Yoneki.

32 records · Page 2Linked to original sources

arXiv2022

BoGraph: Structured Bayesian Optimization From Logs for Expensive Systems with Many Parameters

Current auto-tuning frameworks struggle with tuning computer systems configurations due to their large parameter space, complex interdependencies, and high evaluation cost. Utilizing probabilistic models, Structured Bayesian Optimization (SBO) has recently overcome these difficulties. SBO decomposes the parameter space by utilizing contextual information provided by system experts leading to fast convergence. However, the complexity of building probabilistic models has hindered its wider adoption. We propose BoAnon, a SBO framework that learns the system structure from its logs. BoAnon provides an API enabling experts to encode knowledge of the system as performance models or components dependency. BoAnon takes in the learned structure and transforms it into a probabilistic graph model. Then it applies the expert-provided knowledge to the graph to further contextualize the system behavior. BoAnon probabilistic graph allows the optimizer to find efficient configurations faster than other methods. We evaluate BoAnon via a hardware architecture search problem, achieving an improvement in energy-latency objectives ranging from $5-7$ x-factors improvement over the default architecture. With its novel contextual structure learning pipeline, BoAnon makes using SBO accessible for a wide range of other computer systems such as databases and stream processors.

cs.LG↗

arXiv2021

GDDR: GNN-based Data-Driven Routing

We explore the feasibility of combining Graph Neural Network-based policy architectures with Deep Reinforcement Learning as an approach to problems in systems. This fits particularly well with operations on networks, which naturally take the form of graphs. As a case study, we take the idea of data-driven routing in intradomain traffic engineering, whereby the routing of data in a network can be managed taking into account the data itself. The particular subproblem which we examine is minimising link congestion in networks using knowledge of historic traffic flows. We show through experiments that an approach using Graph Neural Networks (GNNs) performs at least as well as previous work using Multilayer Perceptron architectures. GNNs have the added benefit that they allow for the generalisation of trained agents to different network topologies with no extra work. Furthermore, we believe that this technique is applicable to a far wider selection of problems in systems research.

cs.LG↗

arXiv2021

High-Dimensional Bayesian Optimization with Multi-Task Learning for RocksDB

RocksDB is a general-purpose embedded key-value store used in multiple different settings. Its versatility comes at the cost of complex tuning configurations. This paper investigates maximizing the throughput of RocksDB IO operations by auto-tuning ten parameters of varying ranges. Off-the-shelf optimizers struggle with high-dimensional problem spaces and require a large number of training samples. We propose two techniques to tackle this problem: multi-task modeling and dimensionality reduction through a manual grouping of parameters. By incorporating adjacent optimization in the model, the model converged faster and found complicated settings that other tuners could not find. This approach had an additional computational complexity overhead, which we mitigated by manually assigning parameters to each sub-goal through our knowledge of RocksDB. The model is then incorporated in a standard Bayesian Optimization loop to find parameters that maximize RocksDB's IO throughput. Our method achieved x1.3 improvement when benchmarked against a simulation of Facebook's social graph traffic, and converged in ten optimization steps compared to other state-of-the-art methods that required fifty steps.

cs.LG↗

arXiv2020

The Future is Big Graphs! A Community View on Graph Processing Systems

Graphs are by nature unifying abstractions that can leverage interconnectedness to represent, explore, predict, and explain real- and digital-world phenomena. Although real users and consumers of graph instances and graph workloads understand these abstractions, future problems will require new abstractions and systems. What needs to happen in the next decade for big graph processing to continue to succeed?

cs.DC↗

arXiv2019

Learning Index Selection with Structured Action Spaces

Configuration spaces for computer systems can be challenging for traditional and automatic tuning strategies. Injecting task-specific knowledge into the tuner for a task may allow for more efficient exploration of candidate configurations. We apply this idea to the task of index set selection to accelerate database workloads. Index set selection has been amenable to recent applications of vanilla deep RL, but real deployments remain out of reach. In this paper, we explore how learning index selection can be enhanced with task-specific inductive biases, specifically by encoding these inductive biases in better action structures. Index selection-specific action representations arise when the problem is reformulated in terms of permutation learning and we rely on recent work for learning RL policies on permutations. Through this approach, we build an indexing agent that is able to achieve improved indexing and validate its behavior with task-specific statistics. Early experiments reveal that our agent can find configurations that are up to 40% smaller for the same levels of latency as compared with other approaches and indicate more intuitive indexing behavior.

cs.LG↗

arXiv2019

Wield: Systematic Reinforcement Learning With Progressive Randomization

Reinforcement learning frameworks have introduced abstractions to implement and execute algorithms at scale. They assume standardized simulator interfaces but are not concerned with identifying suitable task representations. We present Wield, a first-of-its kind system to facilitate task design for practical reinforcement learning. Through software primitives, Wield enables practitioners to decouple system-interface and deployment-specific configuration from state and action design. To guide experimentation, Wield further introduces a novel task design protocol and classification scheme centred around staged randomization to incrementally evaluate model capabilities.

cs.LG↗

arXiv2019

RLgraph: Modular Computation Graphs for Deep Reinforcement Learning

Reinforcement learning (RL) tasks are challenging to implement, execute and test due to algorithmic instability, hyper-parameter sensitivity, and heterogeneous distributed communication patterns. We argue for the separation of logical component composition, backend graph definition, and distributed execution. To this end, we introduce RLgraph, a library for designing and executing reinforcement learning tasks in both static graph and define-by-run paradigms. The resulting implementations are robust, incrementally testable, and yield high performance across different deep learning frameworks and distributed backends.

cs.LG↗

arXiv2019

Cloud Futurology

The Cloud has become integral to most Internet-based applications and user gadgets. This article provides a brief history of the Cloud and presents a researcher's view of the prospects for innovating at the infrastructure, middleware, and application and delivery levels of the already crowded Cloud computing stack.

cs.DC↗

arXiv2018

LIFT: Reinforcement Learning in Computer Systems by Learning From Demonstrations

Reinforcement learning approaches have long appealed to the data management community due to their ability to learn to control dynamic behavior from raw system performance. Recent successes in combining deep neural networks with reinforcement learning have sparked significant new interest in this domain. However, practical solutions remain elusive due to large training data requirements, algorithmic instability, and lack of standard tools. In this work, we introduce LIFT, an end-to-end software stack for applying deep reinforcement learning to data management tasks. While prior work has frequently explored applications in simulations, LIFT centers on utilizing human expertise to learn from demonstrations, thus lowering online training times. We further introduce TensorForce, a TensorFlow library for applied deep reinforcement learning exposing a unified declarative interface to common RL algorithms, thus providing a backend to LIFT. We demonstrate the utility of LIFT in two case studies in database compound indexing and resource management in stream processing. Results show LIFT controllers initialized from demonstrations can outperform human baselines and heuristics across latency metrics and space usage by up to 70%.

cs.LG↗

arXiv2016

Tuning the Scheduling of Distributed Stochastic Gradient Descent with Bayesian Optimization

We present an optimizer which uses Bayesian optimization to tune the system parameters of distributed stochastic gradient descent (SGD). Given a specific context, our goal is to quickly find efficient configurations which appropriately balance the load between the available machines to minimize the average SGD iteration time. Our experiments consider setups with over thirty parameters. Traditional Bayesian optimization, which uses a Gaussian process as its model, is not well suited to such high dimensional domains. To reduce convergence time, we exploit the available structure. We design a probabilistic model which simulates the behavior of distributed SGD and use it within Bayesian optimization. Our model can exploit many runtime measurements for inference per evaluation of the objective function. Our experiments show that our resulting optimizer converges to efficient configurations within ten iterations, the optimized configurations outperform those found by generic optimizer in thirty iterations by up to 2X.

stat.ML↗

arXiv2016

Learning Runtime Parameters in Computer Systems with Delayed Experience Injection

Learning effective configurations in computer systems without hand-crafting models for every parameter is a long-standing problem. This paper investigates the use of deep reinforcement learning for runtime parameters of cloud databases under latency constraints. Cloud services serve up to thousands of concurrent requests per second and can adjust critical parameters by leveraging performance metrics. In this work, we use continuous deep reinforcement learning to learn optimal cache expirations for HTTP caching in content delivery networks. To this end, we introduce a technique for asynchronous experience management called delayed experience injection, which facilitates delayed reward and next-state computation in concurrent environments where measurements are not immediately available. Evaluation results show that our approach based on normalized advantage functions and asynchronous CPU-only training outperforms a statistical estimator.

cs.LG↗

arXiv2016

Web Data Knowledge Extraction

A constantly growing amount of information is available through the web. Unfortunately, extracting useful content from this massive amount of data still remains an open issue. The lack of standard data models and structures forces developers to create adhoc solutions from the scratch. The figure of the expert is still needed in many situations where developers do not have the correct background knowledge. This forces developers to spend time acquiring the needed background from the expert. In other directions, there are promising solutions employing machine learning techniques. However, increasing accuracy requires an increase in system complexity that cannot be endured in many projects. In this work, we approach the web knowledge extraction problem using an expertcentric methodology. This methodology defines a set of configurable, extendible and independent components that permit the reutilisation of large pieces of code among projects. Our methodology differs from similar solutions in its expert-driven design. This design, makes it possible for subject-matter expert to drive the knowledge extraction for a given set of documents. Additionally, we propose the utilization of machine assisted solutions that guide the expert during this process. To demonstrate the capabilities of our methodology, we present a real use case scenario in which public procurement data is extracted from the web-based repositories of several public institutions across Europe. We provide insightful details about the challenges we had to deal with in this use case and additional discussions about how to apply our methodology.

cs.DB↗

arXiv2011

On Joint Diagonalisation for Dynamic Network Analysis

Joint diagonalisation (JD) is a technique used to estimate an average eigenspace of a set of matrices. Whilst it has been used successfully in many areas to track the evolution of systems via their eigenvectors; its application in network analysis is novel. The key focus in this paper is the use of JD on matrices of spanning trees of a network. This is especially useful in the case of real-world contact networks in which a single underlying static graph does not exist. The average eigenspace may be used to construct a graph which represents the `average spanning tree' of the network or a representation of the most common propagation paths. We then examine the distribution of deviations from the average and find that this distribution in real-world contact networks is multi-modal; thus indicating several \emph{modes} in the underlying network. These modes are identified and are found to correspond to particular times. Thus JD may be used to decompose the behaviour, in time, of contact networks and produce average static graphs for each time. This may be viewed as a mixture between a dynamic and static graph approach to contact network analysis.

cs.SI↗

arXiv2010

Rhythm and Randomness in Human Contact

There is substantial interest in the effect of human mobility patterns on opportunistic communications. Inspired by recent work revisiting some of the early evidence for a Lévy flight foraging strategy in animals, we analyse datasets on human contact from real world traces. By analysing the distribution of inter-contact times on different time scales and using different graphical forms, we find not only the highly skewed distributions of waiting times highlighted in previous studies but also clear circadian rhythm. The relative visibility of these two components depends strongly on which graphical form is adopted and the range of time scales. We use a simple model to reconstruct the observed behaviour and discuss the implications of this for forwarding efficiency.

physics.data-an↗