# Inductive-bias-driven Reinforcement Learning For Efficient Schedules in Heterogeneous Clusters

arXiv: https://arxiv.org/abs/1909.02119 · HTML: https://mlsanity.com/paper/1909.02119/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Sep 2019.

## Built on by (most cited first, 4 of 4)

- [A Survey on Deep Reinforcement Learning for Data Processing and Analytics](https://mlsanity.com/paper/2108.04526/) (2021)
- [KAIROS: Building Cost-Efficient Machine Learning Inference Systems with Heterogeneous Cloud Resources](https://mlsanity.com/paper/2210.05889/) (2022)
- [BayesPerf: Minimizing Performance Monitoring Errors Using Bayesian Statistics](https://mlsanity.com/paper/2102.10837/) (2021)
- [A Learning Method with Gap-Aware Generation for Heterogeneous DAG Scheduling](https://mlsanity.com/paper/2603.23249/) (2026)

## Abstract (arXiv)

The problem of scheduling of workloads onto heterogeneous processors (e.g., CPUs, GPUs, FPGAs) is of fundamental importance in modern data centers. Current system schedulers rely on application/system-specific heuristics that have to be built on a case-by-case basis. Recent work has demonstrated ML techniques for automating the heuristic search by using black-box approaches which require significant training data and time, which make them challenging to use in practice. This paper presents Symphony, a scheduling framework that addresses the challenge in two ways: (i) a domain-driven Bayesian reinforcement learning (RL) model for scheduling, which inherently models the resource dependencies identified from the system architecture; and (ii) a sampling-based technique to compute the gradients of a Bayesian model without performing full probabilistic inference. Together, these techniques reduce both the amount of training data and the time required to produce scheduling policies that significantly outperform black-box approaches by up to 2.2x.