Two people looking at the exact same scene before them may perceive it differently as a result of a so-called'fingerprint of misperception'. Researchers at the University of California Berkeley found natural variation in the inherent visual ability to pinpoint the exact location and size of objects. A series of experiments on nine individuals found'dramatic differences' in the ability to resolve fine details as well as discrepancies in judging location and size. The differences are due to how the brain processes visual stimuli, the academics believe, but the exact neural network responsible for the variation remains unknown. 'We assume our perception is a perfect reflection of the physical world around us, but this study shows that each of us has a unique visual fingerprint,' study lead author Miss Zixuan Wang, a UC Berkeley doctoral student in psychology, told Berkeley News.

Researchers have created a tool that will make simulations more realistic. A team at the University of California Berkeley used deep reinforcement learning in order to let computer simulations mimic natural human movements. Their tool will allow video game characters to move and animated movie scenes to play out with the fluidity and rhythm of the real world. The recreations of natural movements will make simulations of animals and humans much less clumsy, a report on the new technology said. The feat will even improve scenes that include complex acrobatic feats, such martial arts and break dancing.

To carry out their experiments, the team trained their model in TensorFlow, employing a public dataset of road signs. While the dataset of a few thousand training examples was relatively small, the results plainly show the potential vulnerabilities of deep learning artificial neural networks used in autonomous driving systems when real objects are modified. "Unlike prior work, […] here we focus on evasion attacks where attackers can only modify the testing data instead of training data (poisoning attack)," explained the researchers. "In evasion attacks, an attacker can only change existing physical road signs. Here we assume that an attacker gains access to the classifier after it has been trained ('white-box' access)."

It's very difficult, if not impossible, for us humans to understand how robots see the world. Their cameras work like our eyes do, but the space between the image that a camera captures and actionable information about that image is filled with a black box of machine learning algorithms that are trying to translate patterns of features into something that they're familiar with. Training these algorithms usually involves showing them a set of different pictures of something (like a stop sign), and then seeing if they can extract enough common features from those pictures to reliably identify stop signs that aren't in their training set. This works pretty well, but the common features that machine learning algorithms come up with generally are not "red octagons with the letters S-T-O-P on them." Rather, they're looking features that all stop signs share, but would not be in the least bit comprehensible to a human looking at them.