Open Access. Powered by Scholars. Published by Universities.®

Articles 1 - 5 of 5

Full-Text Articles in Statistical, Nonlinear, and Soft Matter Physics

A Statistical Mechanics Approach To Reinforcement Learning, Jacob Adamczyk Aug 2026

A Statistical Mechanics Approach To Reinforcement Learning, Jacob Adamczyk

Graduate Doctoral Dissertations

Reinforcement learning (RL), the study of optimal decision-making over long timescales in stochastic systems, has recently seen remarkable advances due in large part to the efforts of the deep learning community. RL has witnessed great success in solving problems in video games, robotics, biological control, and language modeling. However, a unified statistical mechanics framework to understand and develop the corresponding algorithms is lacking. To address this issue, we begin by showing that the reinforcement learning problem can be formulated and solved using the tools of statistical mechanics. Drawing on physical principles of free energy minimization and invariance, we address important …


An Application Of The Ising Model, Juliano A. Everett Dec 2019

An Application Of The Ising Model, Juliano A. Everett

Publications and Research

Understanding how the Ising model works,what it represents, and how it can be applied to neurology. Given that an Ising model is an Entropy model that could be representative of the firing of neurons, some assumptions of the system are made and then the process is simulated through Monte Carlo methods.


Universality Class Of Explosive Percolation In Barabási-Albert Networks, Habib E. Islam, M. K. Hassan Jan 2019

Universality Class Of Explosive Percolation In Barabási-Albert Networks, Habib E. Islam, M. K. Hassan

Physics Faculty Publications

In this work, we study explosive percolation (EP) in Barabási-Albert (BA) network, in which nodes are born with degree k = m, for both product rule (PR) and sum rule (SR) of the Achlioptas process. For m = 1 we find that the critical point tc = 1 which is the maximum possible value of the relative link density t; Hence we cannot have access to the other phase like percolation in one dimension. However, for m > 1 we find that tc decreases with increasing m and the critical exponents ν, α, β and γ …


Poincare Recurrence And Spectral Cascades In Three-Dimensional Quantum Turbulence, George Vahala, Jeffrey Yepez, Linda L. Vahala, Min Soe, Bo Zhang, Sean Ziegeler Jan 2011

Poincare Recurrence And Spectral Cascades In Three-Dimensional Quantum Turbulence, George Vahala, Jeffrey Yepez, Linda L. Vahala, Min Soe, Bo Zhang, Sean Ziegeler

Electrical & Computer Engineering Faculty Publications

The time evolution of the ground state wave function of a zero-temperature Bose-Einstein condensate (BEC) gas is well described by the Hamiltonian Gross-Pitaevskii (GP) equation. Using a set of appropriately interleaved unitary collision-stream operators, a qubit lattice gas algorithm is devised, which on taking moments, recovers the Gross-Pitaevskii (GP) equation under diffusion ordering (time scales as length2). Unexpectedly, there is a class of initial states whose Poincaré recurrence time is extremely short and which, as the grid resolution is increased, scales with diffusion ordering (and not as length3). The spectral results of J. Yepez et al. …


Adiabatic Invariance With First Integrals Of Motion, Artur D. Adib Oct 2002

Adiabatic Invariance With First Integrals Of Motion, Artur D. Adib

Dartmouth Scholarship

The construction of a microthermodynamic formalism for isolated systems based on the concept of adiabatic invariance is an old but seldom appreciated effort in the literature, dating back at least to P. Hertz [Ann. Phys. (Leipzig) 33, 225 (1910)]. An apparently independent extension of such formalism for systems bearing additional first integrals of motion was recently proposed by Hans H. Rugh [Phys. Rev. E 64, 055101 (2001)], establishing the concept of adiabatic invariance even in such singular cases. After some remarks in connection with the formalism pioneered by Hertz, it will be suggested that such an extension can …