Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
1
0
1
3
Non-compositionality in machine learning can be systematically repaired through a new categorical framework that reveals how infinitesimal perturbations impact model factorization.