AdvEntuRe: Adversarial Training for Textual Entailment with Knowledge-Guided Examples
This work addresses the challenge of limited training data for textual entailment, which is incremental as it builds on existing methods with novel adversarial techniques.
The paper tackled the problem of learning textual entailment models with limited supervision by proposing knowledge-guided adversarial example generators and a GAN-style training approach, resulting in accuracy improvements of 4.7% on SciTail and 2.8% on a sub-sample of SNLI.
We consider the problem of learning textual entailment models with limited supervision (5K-10K training examples), and present two complementary approaches for it. First, we propose knowledge-guided adversarial example generators for incorporating large lexical resources in entailment models via only a handful of rule templates. Second, to make the entailment model - a discriminator - more robust, we propose the first GAN-style approach for training it using a natural language example generator that iteratively adjusts based on the discriminator's performance. We demonstrate effectiveness using two entailment datasets, where the proposed methods increase accuracy by 4.7% on SciTail and by 2.8% on a 1% training sub-sample of SNLI. Notably, even a single hand-written rule, negate, improves the accuracy on the negation examples in SNLI by 6.1%.