Mitigating attribute amplification in counterfactual image generation

[Free GPT-4]
[DeepSeek]
[PDF] arxiv.org

Robust image representations with counterfactual contrastive learning

M Roschewitz, FDS Ribeiro, T **a, G Khara… - arxiv preprint arxiv …, 2024 - arxiv.org
Contrastive pretraining can substantially increase model generalisation and downstream
performance. However, the quality of the learned representations is highly dependent on the …
保存 引用 関連記事 全 2 バージョン HTMLバージョン