Karthik Yearning Deep Learning

5 Intresting papers from Google AI - Nov01

  1. One Model To Learn Them All

     

    This paper demonstrates a single model to solve problems spanning from multiple domains. This model is trained on ImageNet , COCO dataset , a speech recognition corpus, and an English parsing task.

     

  2. Fluid Annotation

     

    A tool for image annotation. This is a model which performs a strong semantic segmentation, which a human annotator can modify through machine assisted edit operation. Fluid Annotation is a first exploratory step towards making image annotation faster and easier.

     

  3. A Neural Representation of Sketch Drawings

     

    In this paper, a sketch rnn is presented. A recurrent neural network able to construct stroke based drawings of common objects. The team has outlined a framework for conditional and unconditional sketch generation, and describe new robust training methods for generating coherent sketch drawings in a vector format.

     

  4. Neural Architecture Search with Reinforcement Learning

     

    In this paper, the team used a recurrent network to generate the model descriptions of neural networks and train this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set.

     

  5. Adversarial Spheres

     

    This paper is a study of misclassification images by the network which is close to correctly classified images. This is a study to solve the above mentioned adversarial perturbations of the input.