One Search Fits All: Pareto-Optimal Eco-Friendly Model Selection

TL;DR

This paper presents GREEN, an inference-time method for recommending Pareto-optimal AI models that balance validation performance and energy efficiency across multiple domains, addressing limitations of existing eco-efficient search methods.

Contribution

Introduces GREEN, a novel approach using EcoTaskSet dataset and prediction models to recommend energy-efficient AI models tailored to user preferences across diverse tasks.

Findings

Effectively identifies energy-efficient models with competitive performance.

Addresses limitations of existing neural architecture search methods.

Demonstrates applicability across multiple AI domains.

Abstract

The environmental impact of Artificial Intelligence (AI) is emerging as a significant global concern, particularly regarding model training. In this paper, we introduce GREEN (Guided Recommendations of Energy-Efficient Networks), a novel, inference-time approach for recommending Pareto-optimal AI model configurations that optimize validation performance and energy consumption across diverse AI domains and tasks. Our approach directly addresses the limitations of current eco-efficient neural architecture search methods, which are often restricted to specific architectures or tasks. Central to this work is EcoTaskSet, a dataset comprising training dynamics from over 1767 experiments across computer vision, natural language processing, and recommendation systems using both widely used and cutting-edge architectures. Leveraging this dataset and a prediction model, our approach demonstrates effectiveness in selecting the best model configuration based on user preferences. Experimental results show that our method successfully identifies energy-efficient configurations while ensuring competitive performance.