Counterfactual Vision-Language Data Synthesis with Intra-Sample Contrast Learning

Zhecan Wang, Yicheng He, Wenhao Li, Haoxuan You, Long Chen, Noel C Codella, Yulei Niu, Kai-Wei Chang, Shih-Fu Chang

Keywords: counterfactual, data augmentation, vision language, kowledge distillation, vcr, vqa, visual question answering, commonsense reasoning, multimodal, robust, domain-shift, debiased

TL;DR: Counterfactual Vision-Language Data Synthesis with Intra-Sample Contrast Learning for Visual Commonsense Reasoning

Abstract: Existing Visual Learning (VL) benchmarks often contain exploitative biases. Most former works only attempted to mitigate biases in semantically low-level and conventional visual-question-answering typed datasets like VQA and GQA. However, these methods cannot generalize to recently emerging highly semantic VL datasets like VCR and are also difficult to scale due to many severe problems like high-cost labors, drastically disrupting the data distribution\textit{, etc.}To resolve those problems and also address other unique biases on VCR-like datasets, we first conduct in-depth analysis and identify important biases in VCR dataset. We further propose a generalized solution that synthesizes counterfactual image and text data based on the original query's semantic focus while producing less distortion to the data distribution. To utilize our synthesized data, we also design an innovative intra-sample contrastive training strategy to assist QA learning in Visual Commonsense Reasoning (VCR). Moreover, our synthesized VL data also serve as a highly-semantic debiased benchmark for evaluating future VL models' robustness. Extensive experiments show that our proposed synthesized data and training strategy improve existing VL models' performances on both the original VCR dataset and our proposed debiased benchmark.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning