Miles Brundage: "Learning From [Mouse] Brains How to Regularize Machines," Li et al.: https://arxiv.org/abs/1911.05072
Wild stuff - they showed images to mice, recorded the mice's neural activity, made a model of that, then penalized not-mouse-brain-y representations when training new classifiers.
3 replies, 228 likes
Andreas Tolias Lab: Neuroscience has inspired #DeepLearning, but lacks methods to directly translate neural data into better #AI algorithms. Lead by Zhe Li in our @NeurIPSConf paper we used neural data to engineer more robust AI algorithms with better generalization https://arxiv.org/abs/1911.05072.
2 replies, 94 likes
Shahab Bakhtiari: So, if we regularize an ANN to have similar representations as those of mouse visual cortex, the ANN becomes more robust to adversarial attacks.
Here is the most interesting part of the paper for me: https://t.co/EFibwakK5p
0 replies, 22 likes
arxiv: Learning From Brains How to Regularize Machines. http://arxiv.org/abs/1911.05072 https://t.co/X6uY4z3rMD
0 replies, 13 likes
Daisuke Okanohara: Scan neural responses of actual mouses against several images and compute the similarity matrix among images, then use this matrix to regularize NN representation to improve NN's robustness against input noises. Inductive bias from actual brains. https://arxiv.org/abs/1911.05072
0 replies, 12 likes
ALife Papers: Learning From Brains How to Regularize Machines
"We then used the neural representation similarity to regularize CNNs trained on image classification by penalizing intermediate representations that deviated from neural ones"
0 replies, 11 likes
Andreas Tolias Lab: We aim to create AI algorithms that are robust and generalize better using neuroscience data. Learn about our work https://arxiv.org/abs/1911.05072 presented #NeurIPS today (10:45 AM -- 12:45 PM @ East Exhibition Hall B + C #65) and https://www.cell.com/neuron/pdf/S0896-6273(19)30740-8.pdf. Send your CV to email@example.com
0 replies, 8 likes
TübingenNeuroCampus: Cool new paper with contributions from @bethgelab @sinzlab @wielandbr
"Learning From #Brains How to Regularize Machines"
0 replies, 6 likes
Carlos E. Perez 🧢: Well this is certainly wild and a sign of things to come. @Miles_Brundage Research showing using of brain imaging as a regularization parameter for convolutional networks! https://arxiv.org/abs/1911.05072 .
1 replies, 5 likes
Erik Jonker: Some very small but interesting beginnings with regard to improving machine learning with data from natural brains https://arxiv.org/abs/1911.05072
0 replies, 1 likes
Peng Liu: Great work. How about an inverse way? Let’s see “Learning From Machine How to Optimize [Biological] Brains” #AI #DeepLearning #MachineLearning #brain
0 replies, 1 likes
Found on Nov 14 2019 at https://arxiv.org/pdf/1911.05072.pdf