Small LSTMs exhibit capacity-dependent near-critical branching as an emergent regime, with a mixture branching framework proposed to explain 1/f noise under subcritical conditions.
arXiv preprint arXiv:2505.20030 , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
nlin.AO 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Towards Critical Branching Mechanism in Recurrent Neural Networks
Small LSTMs exhibit capacity-dependent near-critical branching as an emergent regime, with a mixture branching framework proposed to explain 1/f noise under subcritical conditions.