Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality

by   Tristan Thrush, et al.

We present a novel task and dataset for evaluating the ability of vision and language models to conduct visio-linguistic compositional reasoning, which we call Winoground. Given two images and two captions, the goal is to match them correctly - but crucially, both captions contain a completely identical set of words, only in a different order. The dataset was carefully hand-curated by expert annotators and is labeled with a rich set of fine-grained tags to assist in analyzing model performance. We probe a diverse range of state-of-the-art vision and language models and find that, surprisingly, none of them do much better than chance. Evidently, these models are not as skilled at visio-linguistic compositional reasoning as we might have hoped. We perform an extensive analysis to obtain insights into how future work might try to mitigate these models' shortcomings. We aim for Winoground to serve as a useful evaluation set for advancing the state of the art and driving further progress in the field. The dataset is available at


page 3

page 8

page 15

page 16

page 17


VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena

We propose VALSE (Vision And Language Structured Evaluation), a novel be...

Compositional Evaluation on Japanese Textual Entailment and Similarity

Natural Language Inference (NLI) and Semantic Textual Similarity (STS) a...

Transformers in the loop: Polarity in neural models of language

Representation of linguistic phenomena in computational language models ...

Analyzing Compositionality-Sensitivity of NLI Models

Success in natural language inference (NLI) should require a model to un...

A Closer Look at Linguistic Knowledge in Masked Language Models: The Case of Relative Clauses in American English

Transformer-based language models achieve high performance on various ta...

Probing via Prompting

Probing is a popular method to discern what linguistic information is co...

Defoiling Foiled Image Captions

We address the task of detecting foiled image captions, i.e. identifying...