Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics
Computation and Language
2021-10-05 v1
Abstract
Much of recent progress in NLU was shown to be due to models' learning dataset-specific heuristics. We conduct a case study of generalization in NLI (from MNLI to the adversarially constructed HANS dataset) in a range of BERT-based architectures (adapters, Siamese Transformers, HEX debiasing), as well as with subsampling the data and increasing the model size. We report 2 successful and 3 unsuccessful strategies, all providing insights into how Transformer-based models learn to generalize.
Cite
@article{arxiv.2110.01518,
title = {Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics},
author = {Prajjwal Bhargava and Aleksandr Drozd and Anna Rogers},
journal= {arXiv preprint arXiv:2110.01518},
year = {2021}
}
Comments
Workshop on Insights from Negative Results (EMNLP 2021)