CIDER (Compactness and Dispersion for OOD detection)

anchor

Representation learning objective for out-of-distribution detection that makes classes compact and pushes their centres apart.

Findings

• IC-028SPADE, an abstaining classifier built on top of ResNet, ViT, and VGG models, detects out-of-distribution and adversarial samples with provable guarantees. [compared-to]
• IC-1405OpenFlamingo and Idefics models hallucinate objects not present in images, and increasing ICL shots beyond 4 amplifies hallucinations [eval]
• IC-1408OpenFlamingo and Idefics models generate low-quality explanations in zero-shot, but ICL and model scale significantly improve explanation CIDEr [eval]
• IC-730CLIPCap and BLIP-2 produce degraded alt-text on Twitter social media images, with BLEU@4 of 0.372 and 0.111 respectively [eval]