CLLGJun 16, 2020

Selective Question Answering under Domain Shift

arXiv:2006.09462v11082 citations
Originality Incremental advance
AI Analysis

This addresses the issue of overconfidence in QA models for users dealing with domain shifts, offering an incremental improvement in abstention policies.

The paper tackles the problem of selective question answering under domain shift, where QA models must abstain from wrong answers when tested on mixed in-domain and out-of-domain data, and proposes a calibrator to predict errors, achieving 56% of questions answered at 80% accuracy compared to 48% with baseline methods.

To avoid giving wrong answers, question answering (QA) models need to know when to abstain from answering. Moreover, users often ask questions that diverge from the model's training data, making errors more likely and thus abstention more critical. In this work, we propose the setting of selective question answering under domain shift, in which a QA model is tested on a mixture of in-domain and out-of-domain data, and must answer (i.e., not abstain on) as many questions as possible while maintaining high accuracy. Abstention policies based solely on the model's softmax probabilities fare poorly, since models are overconfident on out-of-domain inputs. Instead, we train a calibrator to identify inputs on which the QA model errs, and abstain when it predicts an error is likely. Crucially, the calibrator benefits from observing the model's behavior on out-of-domain data, even if from a different domain than the test data. We combine this method with a SQuAD-trained QA model and evaluate on mixtures of SQuAD and five other QA datasets. Our method answers 56% of questions while maintaining 80% accuracy; in contrast, directly using the model's probabilities only answers 48% at 80% accuracy.

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes