ObjectNet is the same size as the ImageNet test set (50,000 images), and by design does not comepaired withatraining setinordertoencourage generalization.

We here argue that popular benchmarks to measure model robustness againstcommon corruptions (likeImageNet-C) underestimate model robustness in many (but not all) application scenarios. The key insight is that inmanyscenarios, multiple unlabeled examples ofthe corruptions are available and can be used for unsupervised online adaptation.

We collect a large real-world test set, ObjectNet, for object recognition with controls where object backgrounds, rotations, and imaging viewpoints are random. Most scientific experiments have controls, confounds which are removed from the data, to ensure that subjects cannot perform a task by exploiting trivial correlations in the data. Historically, large machine learning and computer vision datasets have lacked such controls. This has resulted in models that must be fine-tuned for new datasets and perform better on datasets than in real-world applications. When tested on ObjectNet, object detectors show a 40-45% drop in performance, with respect to their performance on other benchmarks, due to the controls for biases.

Neural Information Processing Systems http://nips.cc/

The prediction updates should improve overall accuracy.