Understanding Out-of-distribution:A Perspective of Data Dynamics

Dyah Adila, Dongyeop Kang

Research output: Contribution to journalConference articlepeer-review

3 Scopus citations

Abstract

Despite machine learning models’ success in Natural Language Processing (NLP) tasks, predictions from these models frequently fail on out-of-distribution (OOD) samples. Prior works have focused on developing state-of-the-art methods for detecting OOD. The fundamental question of how OOD samples differ from in-distribution samples remains unanswered. This paper explores how data dynamics in training models can be used to understand the fundamental differences between OOD and in-distribution samples in extensive detail. We found that syntactic characteristics of the data samples that the model consistently predicts incorrectly in both OOD and in-distribution cases directly contradict each other. In addition, we observed preliminary evidence supporting the hypothesis that models are more likely to latch on trivial syntactic heuristics (e.g., overlap of words between two sentences) when making predictions on OOD samples. We hope our preliminary study accelerates the data-centric analysis on various machine learning phenomena.

Original languageEnglish (US)
Pages (from-to)1-8
Number of pages8
JournalProceedings of Machine Learning Research
Volume163
StatePublished - 2021
Event2nd "I Can't Believe It's Not Better!", ICBINB 2021 Workshop at NeurIPS 2021 - Virtual, Online
Duration: Dec 13 2021 → …

Bibliographical note

Publisher Copyright:
© 2021 Proceedings of Machine Learning Research. All rights reserved.

Fingerprint

Dive into the research topics of 'Understanding Out-of-distribution:A Perspective of Data Dynamics'. Together they form a unique fingerprint.

Cite this