Unveiling the Achilles’ Heel of NLG Evaluators: A Unified Adversarial Framework Driven by Large Language Models
Yiming Chen, Chen Zhang, Danqing Luo, Luis Fernando D’Haro, Robby Tan, Haizhou Li
Abstract
The automatic evaluation of natural language generation (NLG) systems presents a long-lasting challenge. Recent studies have highlighted various neural metrics that align well with human evaluations. Yet, the robustness of these evaluators against adversarial perturbations remains largely under-explored due to the unique challenges in obtaining adversarial data for different NLG evaluation tasks. To address the problem, we introduce AdvEval, a novel black-box adversarial framework against NLG evaluators. AdvEval is specially tailored to generate data that yield strong disagreements between human and victim evaluators. Specifically, inspired by the recent success of large language models (LLMs) in text generation and evaluation, we adopt strong LLMs as both the data generator and gold evaluator. Adversarial data are automatically optimized with feedback from the gold and victim evaluator. We conduct experiments on 12 victim evaluators and 11 NLG datasets, spanning tasks including dialogue, summarization, and question evaluation. The results show that AdvEval can lead to significant performance degradation of various victim metrics, thereby validating its efficacy.- Anthology ID:
- 2024.findings-acl.80
- Volume:
- Findings of the Association for Computational Linguistics: ACL 2024
- Month:
- August
- Year:
- 2024
- Address:
- Bangkok, Thailand
- Editors:
- Lun-Wei Ku, Andre Martins, Vivek Srikumar
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 1359–1375
- Language:
- URL:
- https://aclanthology.org/2024.findings-acl.80
- DOI:
- 10.18653/v1/2024.findings-acl.80
- Cite (ACL):
- Yiming Chen, Chen Zhang, Danqing Luo, Luis Fernando D’Haro, Robby Tan, and Haizhou Li. 2024. Unveiling the Achilles’ Heel of NLG Evaluators: A Unified Adversarial Framework Driven by Large Language Models. In Findings of the Association for Computational Linguistics: ACL 2024, pages 1359–1375, Bangkok, Thailand. Association for Computational Linguistics.
- Cite (Informal):
- Unveiling the Achilles’ Heel of NLG Evaluators: A Unified Adversarial Framework Driven by Large Language Models (Chen et al., Findings 2024)
- PDF:
- https://preview.aclanthology.org/autopr/2024.findings-acl.80.pdf