Recent efforts applying machine learning techniques to query optimization have shown few practical gains due to substantive training overhead, inability to adapt to changes, and poor tail performance. Motivated by these difficulties, we introduce Bao (the \underlineBa ndit \underlineo ptimizer). Bao takes advantage of the wisdom built into existing query optimizers by providing per-query optimization hints. Bao combines modern tree convolutional neural networks with Thompson sampling, a well-studied reinforcement learning algorithm. As a result, Bao automatically learns from its mistakes and adapts to changes in query workloads, data, and schema. Experimentally, we demonstrate that Bao can quickly learn strategies that improve end-to-end query execution performance, including tail latency, for several workloads containing long-running queries. In cloud environments, we show that Bao can offer both reduced costs and better performance compared with a commercial system.

Recently there has been significant interest in using machine learning to improve the accuracy of cardinality estimation. This work has focused on improving average estimation error, but not all estimates matter equally for downstream tasks like query optimization. Since learned models inevitably make mistakes, the goal should be to improve the estimates that make the biggest difference to an optimizer. We introduce a new loss function, Flow-Loss, for learning cardinality estimation models. Flow-Loss approximates the optimizer's cost model and search algorithm with analytical functions, which it uses to optimize explicitly for better query plans. At the heart of Flow-Loss is a reduction of query optimization to a flow routing problem on a certain "plan graph", in which different paths correspond to different query plans. To evaluate our approach, we introduce the Cardinality Estimation Benchmark (CEB) which contains the ground truth cardinalities for sub-plans of over 16 K queries from 21 templates with up to 15 joins. We show that across different architectures and databases, a model trained with Flow-Loss improves the plan costs and query runtimes despite having worse estimation accuracy than a model trained with Q-Error. When the test set queries closely match the training queries, models trained with both loss functions perform well. However, the Q-Error-trained model degrades significantly when evaluated on slightly different queries (e.g., similar but unseen query templates), while the Flow-Loss-trained model generalizes better to such situations, achieving 4 -- 8× better 99th percentile runtimes on unseen templates with the same model architecture and training data.

Contextual policies are used in many settings to customize system parameters and actions to the specifics of a particular setting. In some real-world settings, such as randomized controlled trials or A/B tests, it may not be possible to measure policy outcomes at the level of context—we observe only aggregate rewards across a distribution of contexts. This makes policy optimization much more difficult because we must solve a high-dimensional optimization problem over the entire space of contextual policies, for which existing optimization methods are not suitable. We develop effective models that leverage the structure of the search space to enable contextual policy optimization directly from the aggregate rewards using Bayesian optimization. We use a collection of simulation studies to characterize the performance and robustness of the models, and show that our approach of inferring a low-dimensional context embedding performs best. Finally, we show successful contextual policy optimization in a real-world video bitrate policy problem.

While many deep learning (DL)-based networking systems have demonstrated superior performance, the underlying Deep Neural Networks (DNNs) remain blackboxes and stay uninterpretable for network operators. The lack of interpretability makes DL-based networking systems prohibitive to deploy in practice. In this paper, we propose Metis, a framework that provides interpretability for two general categories of networking problems spanning local and global control. Accordingly, Metis introduces two different interpretation methods based on decision tree and hypergraph, where it converts DNN policies to interpretable rule-based controllers and highlight critical components based on analysis over hypergraph. We evaluate Metis over two categories of state-of-the-art DL-based networking systems and show that Metis provides human-readable interpretations while preserving nearly no degradation in performance. We further present four concrete use cases of Metis, showcasing how Metis helps network operators to design, debug, deploy, and ad-hoc adjust DL-based networking systems.

We present Placeto, a reinforcement learning (RL) approach to efficiently find device placements for distributed neural network training. Unlike prior approaches that only find a device placement for a specific computation graph, Placeto can learn generalizable device placement policies that can be applied to any graph. We propose two key ideas in our approach: (1) we represent the policy as performing iterative placement improvements, rather than outputting a placement in one shot; (2) we use graph embeddings to capture relevant information about the structure of the computation graph, without relying on node labels for indexing. These ideas allow Placeto to train efficiently and generalize to unseen graphs. Our experiments show that Placeto requires up to 6.1x fewer training steps to find placements that are on par with or better than the best placements found by prior approaches. Moreover, Placeto is able to learn a generalizable placement policy for any given family of graphs that can be used without any re-training to predict optimized placements for unseen graphs from the same family. This eliminates the huge overhead incurred by the prior RL approaches whose lack of generalizability necessitates re-training from scratch every time a new graph is to be placed.

We present Park, a platform for researchers to experiment with Reinforcement Learning (RL) for computer systems. Using RL for improving the performance of systems has a lot of potential, but is also in many ways very different from, for example, using RL for games. Thus, in this work we first discuss the unique challenges RL for systems has, and then propose Park an open extensible platform, which makes it easier for ML researchers to work on systems problems. Currently, Park consists of 12 real world system-centric optimization problems with one common easy to use interface. Finally, we present the performance of existing RL approaches over those 12 problems and outline potential areas of future work.

Efficiently scheduling data processing jobs on distributed compute clusters requires complex algorithms. Current systems use simple, generalized heuristics and ignore workload characteristics, since developing and tuning a scheduling policy for each workload is infeasible. In this paper, we show that modern machine learning techniques can generate highly-efficient policies automatically.

Query optimization is one of the most challenging problems in database systems. Despite the progress made over the past decades, query optimizers remain extremely complex components that require a great deal of hand-tuning for specific workloads and datasets. Motivated by this shortcoming and inspired by recent advances in applying machine learning to data management challenges, we introduce Neo ( Neural Optimizer ), a novel learning-based query optimizer that relies on deep neural networks to generate query executions plans. Neo bootstraps its query optimization model from existing optimizers and continues to learn from incoming queries, building upon its successes and learning from its failures. Furthermore, Neo naturally adapts to underlying data patterns and is robust to estimation errors. Experimental results demonstrate that Neo, even when bootstrapped from a simple optimizer like PostgreSQL, can learn a model that offers similar performance to state-of-the-art commercial optimizers, and in some cases even surpass them.

Client-side video players employ adaptive bitrate (ABR) algorithms to optimize user quality of experience (QoE). Despite the abundance of recently proposed schemes, state-of-the-art ABR algorithms suffer from a key limitation: they use fixed control rules based on simplified or inaccurate models of the deployment environment. As a result, existing schemes inevitably fail to achieve optimal performance across a broad set of network conditions and QoE objectives.

Resource management problems in systems and networking often manifest as difficult online decision making tasks where appropriate solutions depend on understanding the workload and environment. Inspired by recent advances in deep reinforcement learning for AI problems, we consider building systems that learn to manage resources directly from experience. We present DeepRM, an example solution that translates the problem of packing tasks with multiple resource demands into a learning problem. Our initial results show that DeepRM performs comparably to state-of-the-art heuristics, adapts to different conditions, converges quickly, and learns strategies that are sensible in hindsight.

We present xFabric, a novel datacenter transport design that provides flexible and fast bandwidth allocation control. xFabric is flexible: it enables operators to specify how bandwidth is allocated amongst contending flows to optimize for different service-level objectives such as minimizing flow completion times, weighted allocations, different notions of fairness, etc. xFabric is also very fast, it converges to the specified allocation one-to-two order of magnitudes faster than prior schemes. Underlying xFabric, is a novel distributed algorithm that uses in-network packet scheduling to rapidly solve general network utility maximization problems for bandwidth allocation. We evaluate xFabric using realistic datacenter topologies and highly dynamic workloads and show that it is able to provide flexibility and fast convergence in such stressful environments.

The evolution of ubiquitous sensing technologies has led to intelligent environments that can monitor and react to our daily activities, such as adapting our heating and cooling systems, responding to our gestures, and monitoring our elderly. In this paper, we ask whether it is possible for smart environments to monitor our vital signs remotely, without instrumenting our bodies. We introduce Vital-Radio, a wireless sensing technology that monitors breathing and heart rate without body contact. Vital-Radio exploits the fact that wireless signals are affected by motion in the environment, including chest movements due to inhaling and exhaling and skin vibrations due to heartbeats. We describe the operation of Vital-Radio and demonstrate through a user study that it can track users' breathing and heart rates with a median accuracy of 99%, even when users are 8~meters away from the device, or in a different room. Furthermore, it can monitor the vital signs of multiple people simultaneously. We envision that Vital-Radio can enable smart homes that monitor people's vital signs without body instrumentation, and actively contribute to their inhabitants' well-being.

We present RF-Capture, a system that captures the human figure -- i.e., a coarse skeleton -- through a wall. RF-Capture tracks the 3D positions of a person's limbs and body parts even when the person is fully occluded from its sensor, and does so without placing any markers on the subject's body. In designing RF-Capture, we built on recent advances in wireless research, which have shown that certain radio frequency (RF) signals can traverse walls and reflect off the human body, allowing for the detection of human motion through walls. In contrast to these past systems which abstract the entire human body as a single point and find the overall location of that point through walls, we show how we can reconstruct various human body parts and stitch them together to capture the human figure. We built a prototype of RF-Capture and tested it on 15 subjects. Our results show that the system can capture a representative human figure through walls and use it to distinguish between various users.

This demo presents Vital-Radio, a wireless sensing technology that monitors breathing remotely, without requiring any body contact. Vital-Radio operates by transmitting a low-power wireless signal and monitoring its reflections off the human body. It uses these reflections to track motion associated with breathing, i.e., the chest movements caused by inhaling and exhaling. The demo will enable any person to sit in front of the device and check that it tracks their inhale and exhale process. The person may hold his/her breath and check that the device detects the breath holding event in real-time.