Tuning and jamming reduced to their minima

Abstract

Inspired by protein folding, we introduce funnels into the complex cost function landscapes of two processes, the tuning of networks, and the jamming of ideal spheres. In both processes, geometrical frustration plays a role -- tuning pressure differences between pairs of target nodes far from the source in a flow network impedes tuning of nearby pairs more than the reverse process, while unjamming the system in one region can make it more difficult to unjam elsewhere. By modifying the cost functions to control the order in which functions are tuned or regions unjam, we smooth out local minima while leaving global minima unaffected, significantly increasing the success rate of reaching global minima.

References in corpus (8)

• Jamming at Zero Temperature and Zero Applied Stress: the Epitome of Disorder
• Gibbs States and the Set of Solutions of Random Constraint Satisfaction Problems
• The jamming transition as a paradigm to understand the loss landscape of deep neural networks
• Multifarious Assembly Mixtures: Systems Allowing Retrieval of Diverse Stored Structures
• The limits of multifunctionality in tunable networks
• Exploring the jamming transition over a wide range of critical densities
• Associative pattern recognition through macro-molecular self-assembly
• Bifurcation analysis and phase diagram of a spin-string model with buckled states

Cited by in corpus (7)

• Demonstration of Decentralized, Physics-Driven Learning
• Supervised learning in physical networks: From machine learning to learning machines
• The Physical Effects of Learning
• Known by the company we keep: `Triadic influence' as a proxy for compatibility in social relationships
• An Analytical Theory of Curriculum Learning in Teacher-Student Networks
• Fluctuation Distributions of Energy Minima in Complex Landscapes
• Physical learning beyond the quasistatic limit