@BigMeanInternet that even if the precise mechanism isn't known. There's also older work showing the earlier vision models see color illusions. For example: https://t.co/T8RNhcEStv https://t.co/mNh2UoagBK

I love this. Although the T800, whose "CPU is a neural net processor...a learning computer," might actually pass this test. ANNs trained on lightness constancy experience lightness illusions too: https://t.co/RaR82qqdYQ

@MIT_CSAIL @davorjord Nice! Reminds me of my much simpler investigation of (shallow) neural nets seeing illusions, from 2007: https://t.co/0xk0U2j1xb

@lottolab I think this goes some way to explain the illusions http://t.co/9PYO5NUW. Also in New Scientist here: http://t.co/wR70EBON