100x Cost & Latency Reduction: Performance Analysis of AI Query Approximation using Lightweight Proxy Models

TL;DR

This paper evaluates a proxy model approach that significantly reduces the cost and latency of AI queries in SQL, maintaining accuracy and enabling scalable, efficient data analysis.

Contribution

It introduces a proxy model-based method for AI query approximation that achieves over 100x cost and latency reduction with preserved accuracy.

Findings

>100x cost and latency reduction for semantic filter operator

Proxy models preserve or improve accuracy across datasets

OLAP and HTAP architectures enable scalable online AI queries

Abstract

Several data warehouse and database providers have recently introduced extensions to SQL called AI Queries, enabling users to specify functions and conditions in SQL that are evaluated by LLMs, thereby broadening significantly the kinds of queries one can express over the combination of structured and unstructured data. LLMs offer remarkable semantic reasoning capabilities, making them an essential tool for complex and nuanced queries that blend structured and unstructured data. While extremely powerful, these AI queries can become prohibitively costly when invoked thousands of times. This paper provides an extensive evaluation of a recent AI query approximation approach that enables low cost analytics and database applications to benefit from AI queries. The approach delivers >100x cost and latency reduction for the semantic filter operator and also important gains for semantic ranking. The cost and performance gains come from utilizing cheap and accurate proxy models over embedding vectors. We show that despite the massive gains in latency and cost, these proxy models preserve accuracy and occasionally improve accuracy across various benchmark datasets, including the extended Amazon reviews benchmark that has 10M rows. We present an OLAP-friendly architecture within Google BigQuery for this approach for purely online (ad hoc) queries, and a low-latency HTAP database-friendly architecture in AlloyDB that could further improve the latency by moving the proxy model training offline. We present techniques that accelerate the proxy model training.