Cutting Down on Prompts and Parameters: Simple Few-Shot Learning with Language Models
This addresses the need for simpler and more efficient few-shot learning methods for NLP practitioners, though it is incremental as it builds on existing finetuning approaches.
The paper tackled the problem of reducing prompt engineering and parameter overhead in few-shot learning with language models by showing that finetuning with null prompts achieves competitive accuracy across tasks, and finetuning only bias terms reduces parameters to 0.1% while maintaining or improving accuracy.
Prompting language models (LMs) with training examples and task descriptions has been seen as critical to recent successes in few-shot learning. In this work, we show that finetuning LMs in the few-shot setting can considerably reduce the need for prompt engineering. In fact, one can use null prompts, prompts that contain neither task-specific templates nor training examples, and achieve competitive accuracy to manually-tuned prompts across a wide range of tasks. While finetuning LMs does introduce new parameters for each downstream task, we show that this memory overhead can be substantially reduced: finetuning only the bias terms can achieve comparable or better accuracy than standard finetuning while only updating 0.1% of the parameters. All in all, we recommend finetuning LMs for few-shot learning as it is more accurate, robust to different prompts, and can be made nearly as efficient as using frozen LMs.