Artificial Intelligence 6.29.2018

The Best Things we read in AI this Week

“Artificial Intelligence” is Vintra’s weekly round-up of AI-related articles, blogs, videos, and papers we liked. 

• Fast and Faster R-CNN … Solving complex computer vision problems. Read the article here.  
  
  

  Everything started with “Rich feature hierarchies for accurate object detection and semantic segmentation” (R-CNN) in 2014, which used an algorithm called Selective Search to propose possible regions of interest and a standard Convolutional Neural Network (CNN) to classify and adjust them. It quickly evolved into Fast R-CNN, published in early 2015, where a technique called Region of Interest Pooling allowed for sharing expensive computations and made the model much faster. Finally came Faster R-CNN, where the first fully differentiable model was proposed.

Check our latest Case Study, The Running Man. 

{{cta(‘c9b7d98e-3cdf-4920-a3cb-44549cbb5b78’)}}

• Towards understanding and eliminating facial bias in AI-powered systems. Read the article here. 
  

  One of the biggest issues causing bias in the area of facial analysis is the lack of diverse data to train systems on. So, this fall, we intend to make publicly available the following dataset as a tool for the technology industry and research community…

• How face recognition software was used to identify the man who shot and killed 5 people at The Capital Gazette. Read the article here. 
  
  

  Police responded to the shooting and say they caught Ramos hiding under a desk, but he had mutilated his fingers in an attempt to avoid identification, a police official confirmed. Another law enforcement source told the Associated Press facial recognition was used to determine Ramos’ identity.

• DeepMind is helping AI imagine scenes it has specifically scene before. Read the article here.  
  

  Given a handful of “snapshots” of a virtual scene, the software—known as a generative query network (GQN)—uses a neural network to build a compact mathematical representation of that scene. It then uses that representation to render images of the room from new perspectives—perspectives the network hasn’t seen before.

Subscribe to our blog to stay up to date!