Assessing Resource-Performance Trade-off of Natural Language Models using Data Envelopment Analysis

TL;DR

This paper introduces a novel application of Data Envelopment Analysis (DEA) to evaluate and compare natural language models based on their resource usage and performance, providing a scalable framework for identifying efficient models.

Contribution

It adapts DEA to assess resource-performance trade-offs in NLP models, offering a new, scalable method to identify models on the efficiency frontier.

Findings

DEA effectively identifies models balancing resources and performance

Empirical analysis on 14 diverse language models demonstrates DEA's utility

DEA reveals a subset of models with optimal resource-performance trade-offs

Abstract

Natural language models are often summarized through a high-dimensional set of descriptive metrics including training corpus size, training time, the number of trainable parameters, inference times, and evaluation statistics that assess performance across tasks. The high dimensional nature of these metrics yields challenges with regard to objectively comparing models; in particular it is challenging to assess the trade-off models make between performance and resources (compute time, memory, etc.). We apply Data Envelopment Analysis (DEA) to this problem of assessing the resource-performance trade-off. DEA is a nonparametric method that measures productive efficiency of abstract units that consume one or more inputs and yield at least one output. We recast natural language models as units suitable for DEA, and we show that DEA can be used to create an effective framework for quantifying model performance and efficiency. A central feature of DEA is that it identifies a subset of models that live on an efficient frontier of performance. DEA is also scalable, having been applied to problems with thousands of units. We report empirical results of DEA applied to 14 different language models that have a variety of architectures, and we show that DEA can be used to identify a subset of models that effectively balance resource demands against performance.