Discover and read the best of Twitter Threads about #GNN
Most recents (3)
"Whole slide images are graphs": Our paper on effectiveness of graph modelling of WSIs in #cpath & Graph Neural Networks (#GNN) for receptor status prediction of #breastcancer from routine WSIs is accepted in Medical Image Analysis.
Preperint: arxiv.org/abs/2110.06042
Preperint: arxiv.org/abs/2110.06042
TL;DR
1. Motivation: Whole slide images are big and holistic modelling of interactions between different tissue components using only slide level labels is difficult with conventional "patch-then-aggregate" approaches used for weakly supervised learning in CPath.
1. Motivation: Whole slide images are big and holistic modelling of interactions between different tissue components using only slide level labels is difficult with conventional "patch-then-aggregate" approaches used for weakly supervised learning in CPath.
2. We introduce a flexible #GNN based #ML framework (#SlideGraph+) that can holistically model WSIs for different large scale prediction problems
3. For HER2 status prediction, the method gives AUC 0.75-0.8 and can be used for case triaging & advanced ordering of tests.
3. For HER2 status prediction, the method gives AUC 0.75-0.8 and can be used for case triaging & advanced ordering of tests.
🚀🚨 Equivariance changes the Scaling Laws of Interatomic Potentials 🚨🚀
We have updated the NequIP paper with up to 3x lower errors + (still) SOTA on MD17
Plus: equivariance changes the power-law exponent of the learning curve!
arxiv.org/abs/2101.03164…
👇🧵 #compchem #GNN
We have updated the NequIP paper with up to 3x lower errors + (still) SOTA on MD17
Plus: equivariance changes the power-law exponent of the learning curve!
arxiv.org/abs/2101.03164…
👇🧵 #compchem #GNN
Learning curves of error vs training set size typically follow a power law: error = a * N^b, where N is the number of training samples and the exponent b determines how fast a method learns as new data become available.
Interestingly, it has been found that different models on the same data set usually only shift the learning curve, but do not change the power-law exponent, see e.g. [1]
[1] arxiv.org/abs/1712.00409
[1] arxiv.org/abs/1712.00409
We are glad to announce that our paper "Message Passing for Hyper-Relational Knowledge Graphs" by @michael_galkin @shape_mismatch @__gauravm @Ricardo_Usbeck @JLehmann82 has been accepted at @emnlp2020! 🎆 We propose a #GNN architecture for hyper-relational KGs like @wikidata. ⬇️
Traditional KGs are based on triples, whereas new KGs like #wikidata use statements and qualifiers to instantiate each edge further making the graph hyper-relational (img1). We incorporate these qualifiers by modifying existing multi-relational GNN (CompGCN) in the StarE (img 2).
This improves link prediction performance across the board!
