小样本学习综述

说到小样本学习一定要先看Meta Learning

小样本学习的名词解释说明

MAML算法提供一个模型无关计算框架，怎么做到模型无关，主要是loss计算不同，计算框架类似adaboost，里面可以换各种算法

Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

原始论文必看

github代码

meta.py的forward可以看出loss的计算

            for k in range(1, self.update_step):
                # 1. run the i-th task and compute loss for k=1~K-1
                logits = self.net(x_spt[i], fast_weights, bn_training=True)
                loss = F.cross_entropy(logits, y_spt[i])
                # 2. compute grad on theta_pi
                grad = torch.autograd.grad(loss, fast_weights)
                # 3. theta_pi = theta_pi - train_lr * grad
                fast_weights = list(map(lambda p: p[1] - self.update_lr * p[0], zip(grad, fast_weights)))

                logits_q = self.net(x_qry[i], fast_weights, bn_training=True)
                # loss_q will be overwritten and just keep the loss_q on last update step.
                loss_q = F.cross_entropy(logits_q, y_qry[i])
                losses_q[k + 1] += loss_q

三种小样本学习方法分类

原型网络讲解

原型网络代码地址

小样本学习在文本分类中的应用

阿里小样本学习翻译

FEW-SHOT TEXT CLASSIFICATION WITH Distributional signature

实践比较好的小样本学习算法

代码见论文中