- https://www.manning.com/books/grokking-deep-learning
- https://www.amazon.com/Adaptive-Filter-Theory-International-Author/dp/B00I6YSQA8/ref=pd_lpo_14_t_1/130-4385160-9989504?_encoding=UTF8&pd_rd_i=B00I6YSQA8&pd_rd_r=3f53b493-1fb0-41d4-ad29-4b6e382a9427&pd_rd_w=PRFO6&pd_rd_wg=gZFTb&pf_rd_p=7b36d496-f366-4631-94d3-61b87b52511b&pf_rd_r=Y8R1318746ECZ880SER2&psc=1&refRID=Y8R1318746ECZ880SER2
- Deep reinforcement learning https://jonathan-hui.medium.com/rl-deep-reinforcement-learning-series-833319a95530
- Self driving car series https://jonathan-hui.medium.com/self-driving-car-series-b8a356f7f2ac
But let’s define some basic concept first.
In CIFAR-10, we have 60,000 images in ten different categories.
In few-shot learning, we train a model with a far larger amount of tasks but far fewer samples per task.
Our ultimate goal is generalizing knowledge and apply the model to new tasks that even not train before.