AI Weekly Update - May 26th, 2021 (#32!)

Thank you for watching! Please subscribe! Content Links: APPS: https://arxiv.org/pdf/2105.09938.pdf Improving Code Autocomplete: https://arxiv.org/pdf/2105.05991.pdf DeepDebug: https://arxiv.org/pdf/2105.09352.pdf The Simplicity Bias: https://arxiv.org/pdf/2105.05612.pdf Rethinking "Batch" in BatchNorm: https://arxiv.org/pdf/2105.07576.pdf Divide and Contrast: https://arxiv.org/pdf/2105.08054.pdf Ethan Perez on True Few-Shot Learning: https://twitter.com/EthanJPerez/status/1397015129506541570 KELM: https://ai.googleblog.com/2021/05/kelm-integrating-knowledge-graphs-with.html Are Larger Pretrained Language Models Uniformly Better? https://arxiv.org/pdf/2105.06020.pdf Pay Attention to MLPs: https://arxiv.org/pdf/2105.08050.pdf Are CNNs or Transformers more like human vision? https://arxiv.org/pdf/2105.07197.pdf FLUTE: https://arxiv.org/pdf/2105.07029.pdf Keras Code Examples - Modern MLP Models: https://keras.io/examples/vision/mlp_image_classification/ Keras Code Examples - Node Classification: https://keras.io/examples/graph/gnn_citations/ BioMed Explorer: https://sites.research.google/biomedexplorer/ BookSum: https://arxiv.org/pdf/2105.08209.pdf Chapters 0:00 Introduction 10:59 APPS 13:27 More Code Deep Learning 15:14 The Simplicity Bias 17:31 Rethinking “Batch” in BatchNorm 20:05 Divide and Contrast 23:01 True Few-Shot Learning 24:20 KELM 25:47 Larger Pretrained Models - Uniformly Better? 26:30 Pay Attention to MLPs 29:47 CNNs or Transformers more like human vision? 31:39 FLUTE 32:51 Pathdreamer 33:50 Keras Code Examples 34:45 Biomedical Explorer 35:23 BookSum