Cross-functional Analysis of Generalisation in Behavioural Learning

TL;DR

This paper introduces BeLUGA, a method for analyzing how well models generalize in behavioural learning across different functionalities, highlighting the risks of overfitting and the effects of various regularization techniques.

Contribution

The paper presents BeLUGA, a novel analysis framework for evaluating behavioural learning and generalization across multiple dimensions and phenomena in NLP tasks.

Findings

BeLUGA effectively measures generalization to unseen functionalities.

Regularisation methods influence the generalization performance.

Models can overfit behavioural test suites, misrepresenting robustness.

Abstract

In behavioural testing, system functionalities underrepresented in the standard evaluation setting (with a held-out test set) are validated through controlled input-output pairs. Optimising performance on the behavioural tests during training (behavioural learning) would improve coverage of phenomena not sufficiently represented in the i.i.d. data and could lead to seemingly more robust models. However, there is the risk that the model narrowly captures spurious correlations from the behavioural test suite, leading to overestimation and misrepresentation of model performance -- one of the original pitfalls of traditional evaluation. In this work, we introduce BeLUGA, an analysis method for evaluating behavioural learning considering generalisation across dimensions of different granularity levels. We optimise behaviour-specific loss functions and evaluate models on several partitions of the behavioural test suite controlled to leave out specific phenomena. An aggregate score measures generalisation to unseen functionalities (or overfitting). We use BeLUGA to examine three representative NLP tasks (sentiment analysis, paraphrase identification and reading comprehension) and compare the impact of a diverse set of regularisation and domain generalisation methods on generalisation performance.