AutoRAG-HP: Automatic Online Hyper-Parameter Tuning for Retrieval-Augmented Generation

TL;DR

AutoRAG-HP introduces an online hyper-parameter tuning framework for Retrieval-Augmented Generation systems, leveraging hierarchical multi-armed bandits to efficiently optimize parameters with fewer API calls.

Contribution

It formulates hyper-parameter tuning as an online MAB problem and proposes a novel Hierarchical MAB method for efficient exploration in RAG systems.

Findings

Achieves Recall@5 ≈ 0.8 with only ~20% of API calls compared to grid search.

Hier-MAB outperforms baselines in challenging optimization scenarios.

Effective tuning of hyper-parameters like top-k, prompt compression, and embedding methods.

Abstract

Recent advancements in Large Language Models have transformed ML/AI development, necessitating a reevaluation of AutoML principles for the Retrieval-Augmented Generation (RAG) systems. To address the challenges of hyper-parameter optimization and online adaptation in RAG, we propose the AutoRAG-HP framework, which formulates the hyper-parameter tuning as an online multi-armed bandit (MAB) problem and introduces a novel two-level Hierarchical MAB (Hier-MAB) method for efficient exploration of large search spaces. We conduct extensive experiments on tuning hyper-parameters, such as top-k retrieved documents, prompt compression ratio, and embedding methods, using the ALCE-ASQA and Natural Questions datasets. Our evaluation from jointly optimization all three hyper-parameters demonstrate that MAB-based online learning methods can achieve Recall@5 $\approx 0.8$ for scenarios with prominent gradients in search space, using only $\sim20\%$ of the LLM API calls required by the Grid Search approach. Additionally, the proposed Hier-MAB approach outperforms other baselines in more challenging optimization scenarios. The code will be made available at https://aka.ms/autorag.