← Search

Naga Sai Abhiram kusumba

2 accepted papers

2025

EraseFlow: Learning Concept Erasure Policies via GFlowNet-Driven Alignment

NeurIPS 2025spotlight

Erasing harmful or proprietary concepts from powerful text‑to‑image generators is an emerging safety requirement, yet current ``concept erasure'' techniques either collapse image quality, rely on brittle adversarial losses, or demand prohibitive retraining cycles. We trace these limitations to a myo…

Cited by 0SourceScholar
2024

TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives

NeurIPS 2024poster

Contrastive Language-Image Pretraining (CLIP) models maximize the mutual information between text and visual modalities to learn representations. This makes the nature of the training data a significant factor in the efficacy of CLIP for downstream tasks. However, the lack of compositional diversity…