Janelle Shane @JanelleCShane Research Scientist in optics. Plays with neural networks. Avid reader, writer, and player of Irish flute. she/her. wandering.shop/@janellecshane Jul. 18, 2019 1 min read

7,500 pictures that image recognition algorithms tend to get wrong. My kind of dataset.

Interesting examples of image recognition algorithms being fooled by color, texture, and background cues.
They also seem to have learned that hummingbird feeder = hummingbird

The image recognition AIs have also learned that bars = rocking chair, shadows+ground = sundial, and thin lines = harvestman spider.
Cruelest of all is the fact that there's no car or van option in ImageNet, but there IS limousine & sports car. Every ride is super fancy.

testing a few examples from the adversarial dataset on visual chatbot. this one worked.

decided to see which "revolver" images were misclassified. apparently ResNet classified this one as "apron".
visual chatbot is similarly unconcerned.

another misclassified "revolver" image

You can follow @JanelleCShane.


Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.