Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Domino: Discovering Systematic Errors with Cross-Modal Embeddings

About

Machine learning models that achieve high overall accuracy often make systematic errors on important subsets (or slices) of data. Identifying underperforming slices is particularly challenging when working with high-dimensional inputs (e.g. images, audio), where important slices are often unlabeled. In order to address this issue, recent studies have proposed automated slice discovery methods (SDMs), which leverage learned model representations to mine input data for slices on which a model performs poorly. To be useful to a practitioner, these methods must identify slices that are both underperforming and coherent (i.e. united by a human-understandable concept). However, no quantitative evaluation framework currently exists for rigorously assessing SDMs with respect to these criteria. Additionally, prior qualitative evaluations have shown that SDMs often identify slices that are incoherent. In this work, we address these challenges by first designing a principled evaluation framework that enables a quantitative comparison of SDMs across 1,235 slice discovery settings in three input domains (natural images, medical images, and time-series data). Then, motivated by the recent development of powerful cross-modal representation learning approaches, we present Domino, an SDM that leverages cross-modal embeddings and a novel error-aware mixture model to discover and describe coherent slices. We find that Domino accurately identifies 36% of the 1,235 slices in our framework - a 12 percentage point improvement over prior methods. Further, Domino is the first SDM that can provide natural language descriptions of identified slices, correctly generating the exact name of the slice in 35% of settings.

Sabri Eyuboglu, Maya Varma, Khaled Saab, Jean-Benoit Delbrouck, Christopher Lee-Messer, Jared Dunnmon, James Zou, Christopher R\'e• 2022

Related benchmarks

TaskDatasetResultRank
Spurious Correlation DiscoveryMNIST, FashionMNIST, and COCO 654 evaluation settings (test)
mP@1017.1
24
Error slice discoveryMetaShift
Precision@1086
17
Error slice discoveryWaterbirds
Precision@1072
17
Error slice discoveryCelebA
Precision@1064
17
Error slice discoveryMNIST-Sum
Precision@1044
17
Subgroup DiscoveryCIFAR-100 1.0 (test)
Cosine Similarity12.2
5
Subgroup DiscoveryBreeds 1.0 (test)
Cosine Similarity10.3
5
Biased subgroup detectionCIFAR-100 (test)
Success Rate35
5
Biased subgroup detectionBreeds (test)
Success Rate38.4
5
Showing 9 of 9 rows

Other info

Follow for update