Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics
Prajjwal Bhargava, Aleksandr Drozd, Anna Rogers
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.
Topics & Concepts
HeuristicsDebiasingGeneralizationTransformerComputer scienceArtificial intelligenceMachine learningGeneralization errorTheoretical computer scienceMathematicsArtificial neural networkCognitive sciencePsychologyEngineeringVoltageOperating systemElectrical engineeringMathematical analysisTopic ModelingNatural Language Processing TechniquesMachine Learning in Healthcare