Unpacking the Resilience of SNLI Contradiction Examples to Attacks

2024-12-15 · Chetan Verma, Archit Agarwal

Pre-trained models excel on NLI benchmarks like SNLI and MultiNLI, but their true language understanding remains uncertain. Models trained only on hypotheses and labels achieve high accuracy, indicating reliance on dataset biases and spurious correlations. To explore this issue, we applied the Universal Adversarial Attack to examine the model's vulnerabilities. Our analysis revealed substantial drops in accuracy for the entailment and neutral classes, whereas the contradiction class exhibited a smaller decline. Fine-tuning the model on an augmented dataset with adversarial examples restored its performance to near-baseline levels for both the standard and challenge sets. Our findings highlight the value of adversarial triggers in identifying spurious correlations and improving robustness while providing insights into the resilience of the contradiction class to adversarial attacks.

📄 PDF Abstract BibTeX arXiv:2412.11172

Code (1)

ckvermaai/snli-attack-analysis 공식 구현 pytorch

Tasks

Adversarial Attack

Teaching a Language Model to Distinguish Between Similar Details using a Small Adversarial Training Set

2024-10-30 · Chris Achard

Language models can achieve high accuracy on natural language tasks such as NLI, but performance suffers on manually created adversarial examples. We investigate the performance of a language model trained on the Stanfor…

Language Modeling Language Modelling Natural Language Inference

First Train to Generate, then Generate to Train: UnitedSynT5 for Few-Shot NLI

2024-12-12 · Sourav Banerjee, Anush Mahajan, Ayushi Agarwal, Eishkaran Singh

Natural Language Inference (NLI) tasks require identifying the relationship between sentence pairs, typically classified as entailment, contradiction, or neutrality. While the current state-of-the-art (SOTA) model, Entai…

Data Augmentation Diversity Few-Shot Learning Few-Shot NLI+3

WikiContradiction: Detecting Self-Contradiction Articles on Wikipedia

2021-11-16 · Cheng Hsu, Cheng-Te Li, Diego Saez-Trumper, Yi-Zhan Hsu

While Wikipedia has been utilized for fact-checking and claim verification to debunk misinformation and disinformation, it is essential to either improve article quality and rule out noisy articles. Self-contradiction is…

Articles Claim Verification Fact Checking Misinformation+1

Asking Crowdworkers to Write Entailment Examples: The Best of Bad Options

2020-10-13 · Asian Chapter of the Association for Computational Linguistics 2020 · Clara Vania, Ruijie Chen, Samuel R. Bowman

Large-scale natural language inference (NLI) datasets such as SNLI or MNLI have been created by asking crowdworkers to read a premise and write three new hypotheses, one for each possible semantic relationships (entailme…

Natural Language Inference Transfer Learning

Delving into adversarial attacks on deep policies

2017-05-18 · Jernej Kos, Dawn Song

Adversarial examples have been shown to exist for a variety of deep learning architectures. Deep reinforcement learning has shown promising results on training agent policies directly on raw inputs such as image pixels. …

Deep Reinforcement Learning reinforcement-learning Reinforcement Learning Reinforcement Learning (RL)