Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics
Abstract
Several strategies, including adapter and debiasing techniques, were tested for their effectiveness in improving the generalization of BERT-based models across different datasets in natural language inference tasks.
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.
Models citing this paper 19
Browse 19 models citing this paperDatasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 13
Collections including this paper 0
No Collection including this paper