Conversational Semantic Role Labeling with Predicate-Oriented Latent Graph

TL;DR

This paper introduces a novel predicate-oriented latent graph approach for conversational semantic role labeling, leveraging a pre-trained dialogue model to improve cross-utterance argument detection in dialogue texts.

Contribution

It proposes a predicate-oriented latent graph induction method and a dialogue-level pre-trained model to enhance CSRL performance, addressing structural and contextual challenges.

Findings

Outperforms baseline models on three CSRL datasets

Achieves over 4% F1 score improvement in cross-utterance argument detection

Demonstrates the effectiveness of predicate-oriented latent graphs and dialogue-level pre-training

Abstract

Conversational semantic role labeling (CSRL) is a newly proposed task that uncovers the shallow semantic structures in a dialogue text. Unfortunately several important characteristics of the CSRL task have been overlooked by the existing works, such as the structural information integration, near-neighbor influence. In this work, we investigate the integration of a latent graph for CSRL. We propose to automatically induce a predicate-oriented latent graph (POLar) with a predicate-centered Gaussian mechanism, by which the nearer and informative words to the predicate will be allocated with more attention. The POLar structure is then dynamically pruned and refined so as to best fit the task need. We additionally introduce an effective dialogue-level pre-trained language model, CoDiaBERT, for better supporting multiple utterance sentences and handling the speaker coreference issue in CSRL. Our system outperforms best-performing baselines on three benchmark CSRL datasets with big margins, especially achieving over 4% F1 score improvements on the cross-utterance argument detection. Further analyses are presented to better understand the effectiveness of our proposed methods.