Opening The Black-Box: Explaining Learned Cost Models For Databases

2025-07-19

AI Summary

Traditional database cost models lack the capability to explain why learned cost models (LCMs)—typically built upon opaque deep neural networks—exhibit large prediction errors for specific query plans, hindering systematic debugging due to their poor interpretability. To address this, we propose the first interpretability framework tailored for database LCMs. Our method integrates structural features of query plans with state-of-the-art deep learning attribution techniques to design a customized attribution algorithm, coupled with an interactive visual analytics tool. This enables fine-grained root-cause localization—such as erroneous operator selection or cardinality estimation bias—and transparently exposes the internal decision logic of LCMs. Experimental evaluation demonstrates that our approach significantly improves error diagnosis efficiency. It provides a reusable, principled methodology to support trustworthy deployment and continuous refinement of learned cost models in modern database systems.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Learned Cost Models (LCMs) have shown superior results over traditional database cost models as they can significantly improve the accuracy of cost predictions. However, LCMs still fail for some query plans, as prediction errors can be large in the tail. Unfortunately, recent LCMs are based on complex deep neural models, and thus, there is no easy way to understand where this accuracy drop is rooted, which critically prevents systematic troubleshooting. In this demo paper, we present the very first approach for opening the black box by bringing AI explainability approaches to LCMs. As a core contribution, we developed new explanation techniques that extend existing methods that are available for the general explainability of AI models and adapt them significantly to be usable for LCMs. In our demo, we provide an interactive tool to showcase how explainability for LCMs works. We believe this is a first step for making LCMs debuggable and thus paving the road for new approaches for systematically fixing problems in LCMs.

Problem

• Explain prediction errors in Learned Cost Models for databases
• Make complex deep neural LCMs interpretable and debuggable
• Adapt AI explainability methods specifically for database LCMs

Innovation

• Extending AI explainability to Learned Cost Models
• Developing tailored explanation techniques for LCMs
• Interactive tool for showcasing LCM explainability