Adversarial Over-Sensitivity and Over-Stability Strategies for Dialogue Models

TL;DR

This paper introduces adversarial strategies to identify and improve the robustness of dialogue models against subtle semantic changes, leading to enhanced performance and state-of-the-art results through adversarial training and subword modeling.

Contribution

It proposes novel adversarial strategies for evaluating and training dialogue models, significantly improving their robustness and performance, and introduces subword-based models for efficiency and resilience.

Findings

Adversarial training improves robustness and performance.

Combined strategies achieve state-of-the-art results.

Subword models are efficient and robust without adversarial training.

Abstract

We present two categories of model-agnostic adversarial strategies that reveal the weaknesses of several generative, task-oriented dialogue models: Should-Not-Change strategies that evaluate over-sensitivity to small and semantics-preserving edits, as well as Should-Change strategies that test if a model is over-stable against subtle yet semantics-changing modifications. We next perform adversarial training with each strategy, employing a max-margin approach for negative generative examples. This not only makes the target dialogue model more robust to the adversarial inputs, but also helps it perform significantly better on the original inputs. Moreover, training on all strategies combined achieves further improvements, achieving a new state-of-the-art performance on the original task (also verified via human evaluation). In addition to adversarial training, we also address the robustness task at the model-level, by feeding it subword units as both inputs and outputs, and show that the resulting model is equally competitive, requires only 1/4 of the original vocabulary size, and is robust to one of the adversarial strategies (to which the original model is vulnerable) even without adversarial training.