Aligning the Pretraining and Finetuning Objectives of Language Models

TL;DR

Aligning pretraining and finetuning objectives in language models enhances performance on specific tasks, especially with limited data, enabling smaller models to achieve high accuracy and reducing labeling costs.

Contribution

This paper introduces the concept of objective alignment during training, demonstrating its effectiveness in improving few-example learning and reducing model size requirements.

Findings

Objective alignment improves task accuracy with fewer finetuning examples.

Small models with aligned objectives outperform larger models without alignment.

Few Example Learning reduces labeling costs and enables real-time applications.

Abstract

We demonstrate that explicitly aligning the pretraining objectives to the finetuning objectives in language model training significantly improves the finetuning task performance and reduces the minimum amount of finetuning examples required. The performance margin gained from objective alignment allows us to build language models with smaller sizes for tasks with less available training data. We provide empirical evidence of these claims by applying objective alignment to concept-of-interest tagging and acronym detection tasks. We found that, with objective alignment, our 768 by 3 and 512 by 3 transformer language models can reach accuracy of 83.9%/82.5% for concept-of-interest tagging and 73.8%/70.2% for acronym detection using only 200 finetuning examples per task, outperforming the 768 by 3 model pretrained without objective alignment by +4.8%/+3.4% and +9.9%/+6.3%. We name finetuning small language models in the presence of hundreds of training examples or less "Few Example learning". In practice, Few Example Learning enabled by objective alignment not only saves human labeling costs, but also makes it possible to leverage language models in more real-time applications.