Sergey Levine    @svlevine    10/11/2021      

In robotic learning, we start from scratch for every experiment, with custom per-task data. Generalization is poor, because the dataset is narrow. In supervised learning domains, there are large reusable datasets (e.g., ImageNet, MS-COCO). What would be "ImageNet" in robotics?
  
    1         8










 
  Related  

Andrew Drozdov    @mrdrozdov    18 hours      

MNIST is really amazing dataset and story. There was a nice talk about its history at ICML in 2016... really wish I could find that link. Also, CIFAR was really important for me in my MS, and Imagenet seems crucial for a large amount of progress with deep learning.
  
    1      



Papers with Datasets    @paperswithdata    12/6/2021      

PartImageNet is a new large dataset with part segmentation annotations. It consists of 158 classes from ImageNet with ~24K images. It can be utilized in multiple vision tasks such as semantic segmentation and few-shot learning. https://t.co/tARLw7XG5A
  
    1         3



Papers with Datasets    @paperswithdata    12/1/2021      

GPR1200 is a new benchmark dataset for general-purpose content-based image retrieval. It enables research to test and evaluate deep learning pre-trained models for generalization qualities. https://t.co/w1OQxLFEUJ
  
    1         1



Chris J. Maddison    @cjmaddison    11/24/2021      

In this talk, I give a simple perspective on representation learning. The recent progress should be blowing your mind: in 2012, a single network (AlexNet) solved a single dataset (ImageNet). In 2021, a single network (CLIP) solved an entire subfield (image classification).
  
    2         8



TheSequence    @TheSequenceAI    11/25/2021      

SwAV is an online algorithm for image classification proposed by @MetaAI researchers. It takes advantage of contrastive learning and avoids comparison between image pairs. SwAV outperforms supervised alternatives and can scale to extremely large datasets.
  
    1         4