Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Affiliation:
2
0
2
7
Most "long-range" GNN benchmarks fail to actually test long-range interactions, but a new framework provides provably long-ranged tasks and a per-range error oracle.
Turns out, the fancy learnable restriction maps in Sheaf Neural Networks might not be worth the complexity – a simple identity mapping works just as well for mitigating oversmoothing on heterophilic graphs.