arXiv subjects

Gal Dalal

Publications and source records attributed to Gal Dalal.

34 records · Page 2Linked to original sources

arXiv2022

Reinforcement Learning for Datacenter Congestion Control

We approach the task of network congestion control in datacenters using Reinforcement Learning (RL). Successful congestion control algorithms can dramatically improve latency and overall network throughput. Until today, no such learning-based algorithms have shown practical potential in this domain. Evidently, the most popular recent deployments rely on rule-based heuristics that are tested on a predetermined set of benchmarks. Consequently, these heuristics do not generalize well to newly-seen scenarios. Contrarily, we devise an RL-based algorithm with the aim of generalizing to different configurations of real-world datacenter networks. We overcome challenges such as partial-observability, non-stationarity, and multi-objectiveness. We further propose a policy gradient algorithm that leverages the analytical structure of the reward function to approximate its derivative and improve stability. We show that this scheme outperforms alternative popular RL approaches, and generalizes to scenarios that were not seen during training. Our experiments, conducted on a realistic simulator that emulates communication networks' behavior, exhibit improved performance concurrently on the multiple considered metrics compared to the popular algorithms deployed today in real datacenters. Our algorithm is being productized to replace heuristics in some of the largest datacenters in the world.

cs.LG↗

arXiv2021

On Covariate Shift of Latent Confounders in Imitation and Reinforcement Learning

We consider the problem of using expert data with unobserved confounders for imitation and reinforcement learning. We begin by defining the problem of learning from confounded expert data in a contextual MDP setup. We analyze the limitations of learning from such data with and without external reward, and propose an adjustment of standard imitation learning algorithms to fit this setup. We then discuss the problem of distribution shift between the expert data and the online environment when the data is only partially observable. We prove possibility and impossibility results for imitation learning under arbitrary distribution shift of the missing covariates. When additional external reward is provided, we propose a sampling procedure that addresses the unknown shift and prove convergence to an optimal solution. Finally, we validate our claims empirically on challenging assistive healthcare and recommender system simulation tasks.

cs.LG↗

arXiv2020

The Architectural Implications of Distributed Reinforcement Learning on CPU-GPU Systems

With deep reinforcement learning (RL) methods achieving results that exceed human capabilities in games, robotics, and simulated environments, continued scaling of RL training is crucial to its deployment in solving complex real-world problems. However, improving the performance scalability and power efficiency of RL training through understanding the architectural implications of CPU-GPU systems remains an open problem. In this work we investigate and improve the performance and power efficiency of distributed RL training on CPU-GPU systems by approaching the problem not solely from the GPU microarchitecture perspective but following a holistic system-level analysis approach. We quantify the overall hardware utilization on a state-of-the-art distributed RL training framework and empirically identify the bottlenecks caused by GPU microarchitectural, algorithmic, and system-level design choices. We show that the GPU microarchitecture itself is well-balanced for state-of-the-art RL frameworks, but further investigation reveals that the number of actors running the environment interactions and the amount of hardware resources available to them are the primary performance and power efficiency limiters. To this end, we introduce a new system design metric, CPU/GPU ratio, and show how to find the optimal balance between CPU and GPU resources when designing scalable and efficient CPU-GPU systems for RL training.

cs.LG↗

arXiv2019

A Tale of Two-Timescale Reinforcement Learning with the Tightest Finite-Time Bound

Policy evaluation in reinforcement learning is often conducted using two-timescale stochastic approximation, which results in various gradient temporal difference methods such as GTD(0), GTD2, and TDC. Here, we provide convergence rate bounds for this suite of algorithms. Algorithms such as these have two iterates, $θ_n$ and $w_n,$ which are updated using two distinct stepsize sequences, $α_n$ and $β_n,$ respectively. Assuming $α_n = n^{-α}$ and $β_n = n^{-β}$ with $1 > α> β> 0,$ we show that, with high probability, the two iterates converge to their respective solutions $θ^*$ and $w^*$ at rates given by $\|θ_n - θ^*\| = \tilde{O}( n^{-α/2})$ and $\|w_n - w^*\| = \tilde{O}(n^{-β/2});$ here, $\tilde{O}$ hides logarithmic terms. Via comparable lower bounds, we show that these bounds are, in fact, tight. To the best of our knowledge, ours is the first finite-time analysis which achieves these rates. While it was known that the two timescale components decouple asymptotically, our results depict this phenomenon more explicitly by showing that it in fact happens from some finite time onwards. Lastly, compared to existing works, our result applies to a broader family of stepsizes, including non-square summable ones.

cs.LG↗

arXiv2019

How to Combine Tree-Search Methods in Reinforcement Learning

Finite-horizon lookahead policies are abundantly used in Reinforcement Learning and demonstrate impressive empirical success. Usually, the lookahead policies are implemented with specific planning methods such as Monte Carlo Tree Search (e.g. in AlphaZero). Referring to the planning problem as tree search, a reasonable practice in these implementations is to back up the value only at the leaves while the information obtained at the root is not leveraged other than for updating the policy. Here, we question the potency of this approach. Namely, the latter procedure is non-contractive in general, and its convergence is not guaranteed. Our proposed enhancement is straightforward and simple: use the return from the optimal tree path to back up the values at the descendants of the root. This leads to a $γ^h$-contracting procedure, where $γ$ is the discount factor and $h$ is the tree depth. To establish our results, we first introduce a notion called \emph{multiple-step greedy consistency}. We then provide convergence rates for two algorithmic instantiations of the above enhancement in the presence of noise injected to both the tree search stage and value estimation stage.

cs.LG↗

arXiv2018

Multiple-Step Greedy Policies in Online and Approximate Reinforcement Learning

Multiple-step lookahead policies have demonstrated high empirical competence in Reinforcement Learning, via the use of Monte Carlo Tree Search or Model Predictive Control. In a recent work \cite{efroni2018beyond}, multiple-step greedy policies and their use in vanilla Policy Iteration algorithms were proposed and analyzed. In this work, we study multiple-step greedy algorithms in more practical setups. We begin by highlighting a counter-intuitive difficulty, arising with soft-policy updates: even in the absence of approximations, and contrary to the 1-step-greedy case, monotonic policy improvement is not guaranteed unless the update stepsize is sufficiently large. Taking particular care about this difficulty, we formulate and analyze online and approximate algorithms that use such a multi-step greedy operator.

cs.LG↗

arXiv2018

Beyond the One Step Greedy Approach in Reinforcement Learning

The famous Policy Iteration algorithm alternates between policy improvement and policy evaluation. Implementations of this algorithm with several variants of the latter evaluation stage, e.g, $n$-step and trace-based returns, have been analyzed in previous works. However, the case of multiple-step lookahead policy improvement, despite the recent increase in empirical evidence of its strength, has to our knowledge not been carefully analyzed yet. In this work, we introduce the first such analysis. Namely, we formulate variants of multiple-step policy improvement, derive new algorithms using these definitions and prove their convergence. Moreover, we show that recent prominent Reinforcement Learning algorithms are, in fact, instances of our framework. We thus shed light on their empirical success and give a recipe for deriving new algorithms for future study.

cs.AI↗

arXiv2018

Finite Sample Analysis of Two-Timescale Stochastic Approximation with Applications to Reinforcement Learning

Two-timescale Stochastic Approximation (SA) algorithms are widely used in Reinforcement Learning (RL). Their iterates have two parts that are updated using distinct stepsizes. In this work, we develop a novel recipe for their finite sample analysis. Using this, we provide a concentration bound, which is the first such result for a two-timescale SA. The type of bound we obtain is known as `lock-in probability'. We also introduce a new projection scheme, in which the time between successive projections increases exponentially. This scheme allows one to elegantly transform a lock-in probability into a convergence rate result for projected two-timescale SA. From this latter result, we then extract key insights on stepsize selection. As an application, we finally obtain convergence rates for the projected two-timescale RL algorithms GTD(0), GTD2, and TDC.

cs.AI↗

arXiv2018

Unit Commitment using Nearest Neighbor as a Short-Term Proxy

We devise the Unit Commitment Nearest Neighbor (UCNN) algorithm to be used as a proxy for quickly approximating outcomes of short-term decisions, to make tractable hierarchical long-term assessment and planning for large power systems. Experimental results on updated versions of IEEE-RTS79 and IEEE-RTS96 show high accuracy measured on operational cost, achieved in runtimes that are lower in several orders of magnitude than the traditional approach.

cs.LG↗

arXiv2018

Safe Exploration in Continuous Action Spaces

We address the problem of deploying a reinforcement learning (RL) agent on a physical system such as a datacenter cooling unit or robot, where critical constraints must never be violated. We show how to exploit the typically smooth dynamics of these systems and enable RL algorithms to never violate constraints during learning. Our technique is to directly add to the policy a safety layer that analytically solves an action correction formulation per each state. The novelty of obtaining an elegant closed-form solution is attained due to a linearized model, learned on past trajectories consisting of arbitrary actions. This is to mimic the real-world circumstances where data logs were generated with a behavior policy that is implausible to describe mathematically; such cases render the known safety-aware off-policy methods inapplicable. We demonstrate the efficacy of our approach on new representative physics-based environments, and prevail where reward shaping fails by maintaining zero constraint violations.

cs.AI↗

arXiv2018

Chance-Constrained Outage Scheduling using a Machine Learning Proxy

Outage scheduling aims at defining, over a horizon of several months to years, when different components needing maintenance should be taken out of operation. Its objective is to minimize operation-cost expectation while satisfying reliability-related constraints. We propose a distributed scenario-based chance-constrained optimization formulation for this problem. To tackle tractability issues arising in large networks, we use machine learning to build a proxy for predicting outcomes of power system operation processes in this context. On the IEEE-RTS79 and IEEE-RTS96 networks, our solution obtains cheaper and more reliable plans than other candidates.

cs.CE↗

arXiv2017

Finite Sample Analyses for TD(0) with Function Approximation

TD(0) is one of the most commonly used algorithms in reinforcement learning. Despite this, there is no existing finite sample analysis for TD(0) with function approximation, even for the linear case. Our work is the first to provide such results. Existing convergence rates for Temporal Difference (TD) methods apply only to somewhat modified versions, e.g., projected variants or ones where stepsizes depend on unknown problem parameters. Our analyses obviate these artificial alterations by exploiting strong properties of TD(0). We provide convergence rates both in expectation and with high-probability. The two are obtained via different approaches that use relatively unknown, recently developed stochastic approximation techniques.

cs.AI↗

arXiv2016

Supervised Learning for Optimal Power Flow as a Real-Time Proxy

In this work we design and compare different supervised learning algorithms to compute the cost of Alternating Current Optimal Power Flow (ACOPF). The motivation for quick calculation of OPF cost outcomes stems from the growing need of algorithmic-based long-term and medium-term planning methodologies in power networks. Integrated in a multiple time-horizon coordination framework, we refer to this approximation module as a proxy for predicting short-term decision outcomes without the need of actual simulation and optimization of them. Our method enables fast approximate calculation of OPF cost with less than 1% error on average, achieved in run-times that are several orders of magnitude lower than of exact computation. Several test-cases such as IEEE-RTS96 are used to demonstrate the efficiency of our approach.

cs.LG↗

arXiv2016

Distributed Scenario-Based Optimization for Asset Management in a Hierarchical Decision Making Environment

Asset management attempts to keep the power system in working conditions. It requires much coordination between multiple entities and long term planning often months in advance. In this work we introduce a mid-term asset management formulation as a stochastic optimization problem, that includes three hierarchical layers of decision making, namely the mid-term, short-term and real-time. We devise a tractable scenario approximation technique for efficiently assessing the complex implications a maintenance schedule inflicts on a power system. This is done using efficient Monte-Carlo simulations that trade-off between accuracy and tractability. We then present our implementation of a distributed scenario-based optimization algorithm for solving our formulation, and use an updated PJM 5-bus system to show a solution that is cheaper than other maintenance heuristics that are likely to be considered by TSOs.

eess.SY↗

arXiv2016

Hierarchical Decision Making In Electricity Grid Management

The power grid is a complex and vital system that necessitates careful reliability management. Managing the grid is a difficult problem with multiple time scales of decision making and stochastic behavior due to renewable energy generations, variable demand and unplanned outages. Solving this problem in the face of uncertainty requires a new methodology with tractable algorithms. In this work, we introduce a new model for hierarchical decision making in complex systems. We apply reinforcement learning (RL) methods to learn a proxy, i.e., a level of abstraction, for real-time power grid reliability. We devise an algorithm that alternates between slow time-scale policy improvement, and fast time-scale value function approximation. We compare our results to prevailing heuristics, and show the strength of our method.

cs.AI↗

arXiv2015

Reinforcement Learning for the Unit Commitment Problem

In this work we solve the day-ahead unit commitment (UC) problem, by formulating it as a Markov decision process (MDP) and finding a low-cost policy for generation scheduling. We present two reinforcement learning algorithms, and devise a third one. We compare our results to previous work that uses simulated annealing (SA), and show a 27% improvement in operation costs, with running time of 2.5 minutes (compared to 2.5 hours of existing state-of-the-art).

cs.AI↗