RSMLP: A light Sampled MLP Structure for Incomplete Utterance Rewrite

TL;DR

This paper introduces RSMLP, a lightweight MLP-based model with a down-sampling strategy for incomplete utterance rewriting, achieving competitive results efficiently.

Contribution

The paper presents a novel, simple, and efficient MLP-based approach with a down-sampling technique for the IUR task, improving performance with minimal complexity.

Findings

Achieves competitive performance on public IUR datasets.

Effective extraction of semantic information between utterances.

Efficient and simple model structure.

Abstract

The Incomplete Utterance Rewriting (IUR) task has garnered significant attention in recent years. Its goal is to reconstruct conversational utterances to better align with the current context, thereby enhancing comprehension. In this paper, we introduce a novel and versatile lightweight method, Rewritten-Sampled MLP (RSMLP). By employing an MLP based architecture with a carefully designed down-sampling strategy, RSMLP effectively extracts latent semantic information between utterances and makes appropriate edits to restore incomplete utterances. Due to its simple yet efficient structure, our method achieves competitive performance on public IUR datasets and in real-world applications.