# Pitfalls and Best Practices in Algorithm Configuration

## TL;DR

This paper discusses common pitfalls in algorithm configuration for AI and proposes best practices and a tool to improve the reliability and effectiveness of automated parameter tuning methods.

## Contribution

It identifies key issues in experimental design for algorithm configuration and introduces GenericWrapper4AC to address these challenges.

## Key findings

- Identified common pitfalls in algorithm configuration experiments.
- Proposed best practices to avoid pitfalls.
- Introduced the GenericWrapper4AC tool for automation.

## Abstract

Good parameter settings are crucial to achieve high performance in many areas of artificial intelligence (AI), such as propositional satisfiability solving, AI planning, scheduling, and machine learning (in particular deep learning). Automated algorithm configuration methods have recently received much attention in the AI community since they replace tedious, irreproducible and error-prone manual parameter tuning and can lead to new state-of-the-art performance. However, practical applications of algorithm configuration are prone to several (often subtle) pitfalls in the experimental design that can render the procedure ineffective. We identify several common issues and propose best practices for avoiding them. As one possibility for automatically handling as many of these as possible, we also propose a tool called GenericWrapper4AC.