Transfer Learning for Performance Modeling of Configurable Systems: A Causal Analysis
This work addresses the challenge of reducing measurement efforts for performance analysis in highly-configurable systems like deep neural networks and big data analytics, representing an incremental advance over previous statistical transfer learning approaches.
The paper tackled the problem of modeling performance in configurable systems by investigating whether causal effects of configuration options can be transferred across environments, confirming that these effects can be carried over with high confidence.
Modern systems (e.g., deep neural networks, big data analytics, and compilers) are highly configurable, which means they expose different performance behavior under different configurations. The fundamental challenge is that one cannot simply measure all configurations due to the sheer size of the configuration space. Transfer learning has been used to reduce the measurement efforts by transferring knowledge about performance behavior of systems across environments. Previously, research has shown that statistical models are indeed transferable across environments. In this work, we investigate identifiability and transportability of causal effects and statistical relations in highly-configurable systems. Our causal analysis agrees with previous exploratory analysis \cite{Jamshidi17} and confirms that the causal effects of configuration options can be carried over across environments with high confidence. We expect that the ability to carry over causal relations will enable effective performance analysis of highly-configurable systems.