Heterophilous graphs, in which edges connect dissimilar nodes, are challenging for GNNs. In our #iclr2023 paper, we propose a new benchmark of 5 structurally diverse heterophilous graph datasets. Read more in this post by Oleg Platonov and @LProkhorenkova https://t.co/JI8TPYy0f4
We present SWARM, an efficient algorithm for model-parallel training across the Internet (e.g. with volunteers).
Key advantages:
💎 Fault-tolerant
⚖️ Self-balancing on slow GPUs/networks
🐌 Works in low-bandwidth setups
📜 https://t.co/wCnf6vDCv4
🖥️ https://t.co/pVe0GDmfrK
@PeterZhizhin Гипотеза: нахождение в стране по визе привязанной к работе может считаться поводом не сокращать в первую очередь. Думаю, эту идею возможно донести до этого вашего совета.
@peter_richtarik@iclr_conf@icmlconf@NeurIPSConf Perhaps it is worth looking at the other side, and somehow support papers that "scale up and show to work on real problems" too? Because now they are rejected due to "lack of novelty".