技术标签:Domain Adaptation计算机视觉 论文地址:https://ieeexplore.ieee.org/document/9010003 源码地址:https://github.com/He-Zhenwei/MAF 1 Core Idea 多对抗Faster RCNN潜在地解决了域自适应在特征上的不一致,本文的方法包括三部分:1)分层域特征对齐模块;2)信息不变性尺度降低模块(information invariant scale ...
Semi-Supervised Domain Adaptation Poster 2562 An Ensemble of Epoch-wise Empirical Bayes for Few-shot 【ICLR2019】Poster 论文汇总 current 3D learning literature. Augmented Cyclic Adversarial Learning for Low Resource Domain... MAML Keywords:meta-learning, deep-learning, few-shot learning, supervised ...