Paper Reviews

Excessive Invariance Causes Adversarial Vulnerability
On The Power of Curriculum Learning in Training Deep Networks
Learning Transferable Visual Models From Natural Language Supervision
ConvNets and ImageNet Beyond Accuracy: Understanding Mistakes and Uncovering Biases
Could a Neuroscientist Understand a Microprocessor?
Humans, but Not Deep Neural Networks, Often Miss Giant Targets in Scenes
Embodied Intelligence via Learning and Evolution
Training data-efficient image transformers & distillation through attention