3rd May - 16th May 2019

Etymo Newsletter provides the latest development in machine learning research, including the most popular datasets and the most trending papers in the past two weeks.

If you like this newsletter, you can subscribe to our fortnightly newsletters here.

Fortnight Summary

Popular datasets include MNIST, ImageNet, CIFAR-10, COCO, and KITTI. Our trending phrases are Random Gradient, Signal Control, and Market Price. Three trending papers are related to semi-supervised learning, adversarial examples, and similarity index of neural network representations. We also present you a couple of papers focusing on the applications of Machine Learning.

Popular Datasets

Here are most mentioned datasets over the last two weeks.

Name Type Number of Papers
MNIST Handwritten Digits 73
ImageNet Image Dataset 56
CIFAR-10 Tiny Image Dataset in 10 Classes 47
COCO Common Objects in Context 27
KITTI Autonomous Driving 22

Trending Phrases

In this section, we present a list of phrases that appeared significantly more in this newsletter than the previous newsletters.

Etymo Trending

Presented below is a list of the most trending papers added in the last two weeks.

  • MixMatch: A Holistic Approach to Semi-Supervised Learning:
    MixMatch is a unified method of the current dominant approaches for semi-supervised learning. It guesses low-entropy labels for data-augmented unlabeled examples and mixes labeled and unlabeled data using MixUp. MixMatch obtains state-of-the-art results by a large margin across many datasets and labeled data amounts, and achieves a dramatically better accuracy-privacy trade-off for differential privacy.

  • Adversarial Examples Are Not Bugs, They Are Features:
    Adversarial examples can be directly attributed to the presence of non-robust features: features derived from patterns in the data distribution that are highly predictive, yet brittle and incomprehensible to humans. After capturing these features within a theoretical framework, the authors establish their widespread existence in standard datasets.

  • Similarity of Neural Network Representations Revisited:
    The authors show that canonical correlation analysis (CCA) belongs to a family of statistics for measuring multivariate similarity, but that neither CCA nor any other statistic that is invariant to invertible linear transformation can measure meaningful similarities between representations of higher dimension than the number of data points. They introduce a similarity index that measures the relationship between representational similarity matrices and does not suffer from this limitation.

Papers on Machine Learning Application

  • Traffic Queue Length and Pressure Estimation for Road Networks with Geometric Deep Learning Algorithms:
    In this paper, relatively inexpensive and easy to install loop-detectors are used by a geometric deep learning algorithm, which uses loop-detector data in a spatial context of a road network, to estimate queue length in front of signalized intersections. The data then can be used for following traffic control tasks.
  • Deep Neural Networks for Marine Debris Detection in Sonar Images:
    Marine debris can pose significant ecological problems. Better marine debris detection can enhance the better collection from underwater envrionments. This paper performs a comprehensive evaluation on the use of DNNs for the problem of marine debris detection in FLS images, as well as related problems such as image classification, matching, and detection proposals.

Hope you have enjoyed this newsletter! If you have any comments or suggestions, please email or