GenSF: Simultaneous Adaptation of Generative Pre-trained Models and Slot Filling

TL;DR

GenSF introduces a novel method for slot filling that simultaneously adapts pre-trained models and reformulates downstream tasks, achieving state-of-the-art results especially in few-shot and zero-shot scenarios.

Contribution

It proposes a scalable approach that aligns pre-trained models with downstream tasks by joint adaptation, avoiding task-specific pre-training objectives.

Findings

9 F1 score improvement in zero-shot slot filling

State-of-the-art results on two datasets

Strong gains in few-shot and zero-shot settings

Abstract

In transfer learning, it is imperative to achieve strong alignment between a pre-trained model and a downstream task. Prior work has done this by proposing task-specific pre-training objectives, which sacrifices the inherent scalability of the transfer learning paradigm. We instead achieve strong alignment by simultaneously modifying both the pre-trained model and the formulation of the downstream task, which is more efficient and preserves the scalability of transfer learning. We present GenSF (Generative Slot Filling), which leverages a generative pre-trained open-domain dialog model for slot filling. GenSF (1) adapts the pre-trained model by incorporating inductive biases about the task and (2) adapts the downstream task by reformulating slot filling to better leverage the pre-trained model's capabilities. GenSF achieves state-of-the-art results on two slot filling datasets with strong gains in few-shot and zero-shot settings. We achieve a 9 F1 score improvement in zero-shot slot filling. This highlights the value of strong alignment between the pre-trained model and the downstream task.