Mass conservation promotes self-organized criticality in neural automata reservoirs

Researchers have explored the use of mass conservation as an inductive bias to promote self-organized criticality (SOC) in neural cellular automata (NCA) reservoirs. This approach, which involves a local redistribution rule that preserves total lattice mass, was found to consistently exhibit stronger criticality compared to standard NCAs. The mass-conserving NCAs were also faster to evolve and achieved comparable performance on downstream tasks such as sequential memory, digit classification, and temporal control, suggesting that mass conservation is an effective mechanism for enhancing criticality without compromising utility. AI

IMPACT This research could lead to more efficient and effective neural network architectures for tasks requiring temporal processing and memory.

RANK_REASON The cluster contains a research paper detailing a novel approach to improving neural automata reservoirs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.NE (Neural & Evolutionary) →

• 5-bit sequential memory
• CartPole-v1
• MNIST
• NCA Reservoirs
• Neural Cellular Automata
• Self-Organized Criticality
• Tong Zhang

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →