Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Networks 閱讀筆記

Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Networks 收穫 多個網絡如何組合? 相同目標collaborative, 不同目標adversarial todo 查看groud truth class embedding 查看SAE Conclusion pro using adv

>>阅读原文<<