Search papers, labs, and topics across Lattice.
Affiliation:
4
0
5
27
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.
Standard port-Hamiltonian neural networks are fundamentally limited to single-attractor systems; this work unlocks multi-attractor dynamics by novelly parametrizing the Hamiltonian as a product of Bregman divergences.
Adversarial examples can exploit reconstruction-based detectors, leading to a significant drop in detection accuracy even under real-world conditions.
Not all layers are created equal: pruning the KV cache in a layer-dependent manner significantly boosts long-context LLM performance compared to uniform pruning strategies.
