Accelerating Quadratic Optimization with Reinforcement Learning

First-order methods for quadratic optimization such as OSQP are widely used for large-scale machine learning and embedded optimal control, where many related problems must be rapidly solved. These methods face two persistent challenges: manual hyperparameter tuning and convergence time to high-accuracy solutions. To address these, we explore how Reinforcement Learning (RL) can learn a policy to tune parameters to accelerate convergence. In experiments with well-known QP benchmarks we find that our RL policy, RLQP, significantly outperforms state-of-the-art QP solvers by up to 3x. RLQP generalizes surprisingly well to previously unseen problems with varying dimension and structure from different applications, including the QPLIB, Netlib LP and Maros-Meszaros problems. Code for RLQP is available at https://github.com/berkeleyautomation/rlqp.

Scenic4RL: Programmatic Modeling and Generation of Reinforcement Learning Environments

The capability of reinforcement learning (RL) agent directly depends on the diversity of learning scenarios the environment generates and how closely it captures real-world situations. However, existing environments/simulators lack the support to systematically model distributions over initial states and transition dynamics. Furthermore, in complex domains such as soccer, the space of possible scenarios is infinite, which makes it impossible for one research group to provide a comprehensive set of scenarios to train, test, and benchmark RL algorithms. To address this issue, for the first time, we adopt an existing formal scenario specification language, SCENIC, to intuitively model and generate interactive scenarios. We interfaced SCENIC to Google Research Soccer environment to create a platform called SCENIC4RL. Using this platform, we provide a dataset consisting of 36 scenario programs encoded in SCENIC and demonstration data generated from a subset of them. We share our experimental results to show the effectiveness of our dataset and the platform to train, test, and benchmark RL algorithms. More importantly, we open-source our platform to enable RL community to collectively contribute to constructing a comprehensive set of scenarios.

Online Learning of Competitive Equilibria in Exchange Economies

The sharing of scarce resources among multiple rational agents is one of the classical problems in economics. In exchange economies, which are used to model such situations, agents begin with an initial endowment of resources and exchange them in a way that is mutually beneficial until they reach a competitive equilibrium (CE). CE allocations are Pareto efficient and fair. Consequently, they are used widely in designing mechanisms for fair division. However, computing CEs requires the knowledge of agent preferences which are unknown in several applications of interest. In this work, we explore a new online learning mechanism, which, on each round, allocates resources to the agents and collects stochastic feedback on their experience in using that allocation. Its goal is to learn the agent utilities via this feedback and imitate the allocations at a CE in the long run. We quantify CE behavior via two losses and propose a randomized algorithm which achieves $\bigOtilde(\sqrt{T})$ loss after $T$ rounds under both criteria. Empirically, we demonstrate the effectiveness of this mechanism through numerical simulations.

PAC Best Arm Identification Under a Deadline

We study $(\epsilon, \delta)$-PAC best arm identification, where a decision-maker must identify an $\epsilon$-optimal arm with probability at least $1 - \delta$, while minimizing the number of arm pulls (samples). Most of the work on this topic is in the sequential setting, where there is no constraint on the time taken to identify such an arm; this allows the decision-maker to pull one arm at a time. In this work, the decision-maker is given a deadline of $T$ rounds, where, on each round, it can adaptively choose which arms to pull and how many times to pull them; this distinguishes the number of decisions made (i.e., time or number of rounds) from the number of samples acquired (cost). Such situations occur in clinical trials, where one may need to identify a promising treatment under a deadline while minimizing the number of test subjects, or in simulation-based studies run on the cloud, where we can elastically scale up or down the number of virtual machines to conduct as many experiments as we wish, but need to pay for the resource-time used. As the decision-maker can only make $T$ decisions, she may need to pull some arms excessively relative to a sequential algorithm in order to perform well on all possible problems. We formalize this added difficulty with two hardness results that indicate that unlike sequential settings, the ability to adapt to the problem difficulty is constrained by the finite deadline. We propose Elastic Batch Racing (EBR), a novel algorithm for this setting and bound its sample complexity, showing that EBR is optimal with respect to both hardness results. We present simulations evaluating EBR in this setting, where it outperforms baselines by several orders of magnitude.

ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training

The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for training neural networks within limited memory environments. In this work, we propose ActNN, a memory-efficient training framework that stores randomly quantized activations for back propagation. We prove the convergence of ActNN for general network architectures, and we characterize the impact of quantization on the convergence via an exact expression for the gradient variance. Using our theory, we propose novel mixed-precision quantization strategies that exploit the activation's heterogeneity across feature dimensions, samples, and layers. These techniques can be readily applied to existing dynamic graph frameworks, such as PyTorch, simply by substituting the layers. We evaluate ActNN on mainstream computer vision models for classification, detection, and segmentation tasks. On all these tasks, ActNN compresses the activation to 2 bits on average, with negligible accuracy loss. ActNN reduces the memory footprint of the activation by 12x, and it enables training with a 6.6x to 14x larger batch size.

Pylot: A Modular Platform for Exploring Latency-Accuracy Tradeoffs in Autonomous Vehicles

We present Pylot, a platform for autonomous vehicle (AV) research and development, built with the goal to allow researchers to study the effects of the latency and accuracy of their models and algorithms on the end-to-end driving behavior of an AV. This is achieved through a modular structure enabled by our high-performance dataflow system that represents AV software pipeline components (object detectors, motion planners, etc.) as a dataflow graph of operators which communicate on data streams using timestamped messages. Pylot readily interfaces with popular AV simulators like CARLA, and is easily deployable to real-world vehicles with minimal code changes. To reduce the burden of developing an entire pipeline for evaluating a single component, Pylot provides several state-of-the-art reference implementations for the various components of an AV pipeline. Using these reference implementations, a Pylot-based AV pipeline is able to drive a real vehicle, and attains a high score on the CARLA Autonomous Driving Challenge. We also present several case studies enabled by Pylot, including evidence of a need for context-dependent components, and per-component time allocation. Pylot is open source, with the code available at https://github.com/erdos-project/pylot.

TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models

Model parallelism has become a necessity for training modern large-scale deep language models. In this work, we identify a new and orthogonal dimension from existing model parallel approaches: it is possible to perform pipeline parallelism within a single training sequence for Transformer-based language models thanks to its autoregressive property. This enables a more fine-grained pipeline compared with previous work. With this key idea, we design TeraPipe, a high-performance token-level pipeline parallel algorithm for synchronous model-parallel training of Transformer-based language models. We develop a novel dynamic programming-based algorithm to calculate the optimal pipelining execution scheme given a specific model and cluster configuration. We show that TeraPipe can speed up the training by 5.0x for the largest GPT-3 model with 175 billion parameters on an AWS cluster with 48 p3.16xlarge instances compared with state-of-the-art model-parallel methods.

Ekya: Continuous Learning of Video Analytics Models on Edge Compute Servers

Video analytics applications use edge compute servers for the analytics of the videos (for bandwidth and privacy). Compressed models that are deployed on the edge servers for inference suffer from data drift, where the live video data diverges from the training data. Continuous learning handles data drift by periodically retraining the models on new data. Our work addresses the challenge of jointly supporting inference and retraining tasks on edge servers, which requires navigating the fundamental tradeoff between the retrained model's accuracy and the inference accuracy. Our solution Ekya balances this tradeoff across multiple models and uses a micro-profiler to identify the models that will benefit the most by retraining. Ekya's accuracy gain compared to a baseline scheduler is 29% higher, and the baseline requires 4x more GPU resources to achieve the same accuracy as Ekya.

Online Learning Demands in Max-min Fairness

We describe mechanisms for the allocation of a scarce resource among multiple users in a way that is efficient, fair, and strategy-proof, but when users do not know their resource requirements. The mechanism is repeated for multiple rounds and a user's requirements can change on each round. At the end of each round, users provide feedback about the allocation they received, enabling the mechanism to learn user preferences over time. Such situations are common in the shared usage of a compute cluster among many users in an organisation, where all teams may not precisely know the amount of resources needed to execute their jobs. By understating their requirements, users will receive less than they need and consequently not achieve their goals. By overstating them, they may siphon away precious resources that could be useful to others in the organisation. We formalise this task of online learning in fair division via notions of efficiency, fairness, and strategy-proofness applicable to this setting, and study this problem under three types of feedback: when the users' observations are deterministic, when they are stochastic and follow a parametric model, and when they are stochastic and nonparametric. We derive mechanisms inspired by the classical max-min fairness procedure that achieve these requisites, and quantify the extent to which they are achieved via asymptotic rates. We corroborate these insights with an experimental evaluation on synthetic problems and a web-serving task.