Computer Vision News - September 2018

Smaller Collaborative Robots Disrupting Robotics Industry: Very complete review of current issues and solutions to robots in the industry and how a new category of robots is rapidly becoming mainstream - here is how RSIP Vision does it . BTW, if you want to see how a robot learns dexterity, Watch the Video on the Right . Top 23 Deep Learning Papers: Why exactly 23 papers I don’t know: you should ask the guys at marktechpost.com who prepared the list. It is not new but we noticed it only very recently. Regardless, it is a remarkable choice of Deep Learning papers . Among the authors: Jürgen Schmidhuber , Yoshua Bengio , Yann LeCun , Matt Zeiler , Kaiming He , Lisa Anne Hendricks , Kate Saenko , Tamara Berg , Subhashini Venugopalan , Marcus Rohrbach , and many more friends of Computer Vision News. The 23 Papers Super VIP Cheatsheet: Machine Learning and others! Let’s start from the top of the tops: a 15-pages cheatsheet covering the content of Stanford's CS229 class on Machine Learning , with content originally taught by Fei Fei Li , Andrew NG and others. More cheatsheets: Deep Learning , Supervised Learning , Unsupervised Learning , Tips and tricks . Pure Gold! Why computers are so bad at comparing objects: Very clear article explaining in simple terms why computers are so bad at a class of tasks that almost no human has problem with : determining whether two objects in an image are the same or different. Read it Now! 9 Things You Should Know About TensorFlow: Last month many readers liked the piece by Cassie Kozyrkov . So here we found a new one, even more interesting, reviewing a talk on Google Cloud Next about what’s new with TensorFlow . Cassie’s great article is here and La uren ce Moroney’s video is here . Spotlight News 11 Computer Vision News Artificial Intelligence Draws Faces from Text Description s: It’s a brilliant idea: it tries to give a face to book characters we can only figure in our mind or watch in movies. His project called T2F uses a Generative Adversarial Network (GAN) to encode text and synthesize facial images. Watch the Video on the Left .