No idea how to implement this..but looks interesting
New code example on http://keras.io : a message-passing neural network for predicting properties of molecules given their graph. https://keras.io/examples/graph/mpnn-molecular-graphs/ …pic.twitter.com/rdDs6pRYe6
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is what a lot of the focus has been on in computational biology with artificial intelligence. Given the structure of a protein, how could we predict its function? Very interesting that we have this NN in Keras now.
#Keras#TensorFlow#MachineLearning#DeepLearningThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This guide is AMAZING! This guide was the missing info I needed to make the encoder-process-decoder GNN architecture (deepmind). Two main take-aways: 1. tf.dataset shaneningans for graphs 2. tf.gather and message-passing
-
Here is a cool looking result of shortest path prediction on graphs using one hot encoding made from scratch (in tensorflow)using the aforementioned EncoderProcessDecode archit. I cannot thank the community enough !! (later this month I will upload a collab for anyone intrested)pic.twitter.com/UBLMM3QkOI
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.