Multi-Prompting Decoder Helps Better Language Understanding

TL;DR

This paper introduces a Multi-Prompting Decoder framework that queries pre-trained language models with multiple prompts, improving natural language understanding performance especially in few-shot scenarios by reducing prompt dependency and extracting richer knowledge.

Contribution

The paper proposes a novel multi-prompting decoding approach for Model-as-a-Service PLMs, enhancing task adaptation and performance in few-shot settings.

Findings

Achieves new state-of-the-art results on multiple NLU datasets.

Mitigates reliance on prompt quality and alleviates data scarcity issues.

Effective in few-shot learning scenarios.

Abstract

Recent Pre-trained Language Models (PLMs) usually only provide users with the inference APIs, namely the emerging Model-as-a-Service (MaaS) setting. To adapt MaaS PLMs to downstream tasks without accessing their parameters and gradients, some existing methods focus on the output-side adaptation of PLMs, viewing the PLM as an encoder and then optimizing a task-specific decoder for decoding the output hidden states and class scores of the PLM. Despite the effectiveness of these methods, they only use a single prompt to query PLMs for decoding, leading to a heavy reliance on the quality of the adopted prompt. In this paper, we propose a simple yet effective Multi-Prompting Decoder (MPD) framework for MaaS adaptation. The core idea is to query PLMs with multiple different prompts for each sample, thereby obtaining multiple output hidden states and class scores for subsequent decoding. Such multi-prompting decoding paradigm can simultaneously mitigate reliance on the quality of a single prompt, alleviate the issue of data scarcity under the few-shot setting, and provide richer knowledge extracted from PLMs. Specifically, we propose two decoding strategies: multi-prompting decoding with optimal transport for hidden states and calibrated decoding for class scores. Extensive experiments demonstrate that our method achieves new state-of-the-art results on multiple natural language understanding datasets under the few-shot setting.