1. Vision & Multimodal

- Visual Structure: scenes as objects, parts, and motion

InstaGAN (ICLR'19), OACon (NeurIPS'21), OAMixer (CVPRW'22), DIGAN (ICLR'22),

CAST (ICLR'24), CAFT (NeurIPS'26), SHED (TMLR'26)

- Learning in the Wild: limited, uncurated, or novel data

CSI (NeurIPS'20), MASKER (AAAI'21), RoPAWS (ICLR'23), S-CLIP (NeurIPS'23),

OAK (CVPR'25), ToMA (NeurIPS'26), T-PRH (arXiv'26)

- Reliable Generation: under bias, noise, and scarce data

GOLD (NeurIPS'19), FreezeD (CVPRW'20), FICS (TMLR'23), StayFair (NeurIPS'26)