Hello Everyone,
Tomorrow I'll be giving my first group meeting at the usual time, in
the usual place (Division Room, 2pm). My plan is to present two
projects that I've been working on recently; thus, I'll include two
abstracts.
Part 1: Effective classical force-fields for quantum ensembles
Most approaches to molecular dynamics involve integrating Newton's
equations on force-fields which are designed to approximate quantum
potential energy surfaces. While this seems justifiable in the limit
of high temperature and/or heavy atoms, the "highly quantum" behavior
of ensembles significantly affected by low mass particles such as
hydrogen cannot be reproduced via classical propagation with such
force-fields. However, the success of path integral approaches such as
RPMD beg the question of whether there may exist local "effective
potentials" which can approximately reproduce quantum ensemble
distributions when integrated classically. Using a simple genetic
algorithm routine we have been able to find effective potentials which
closely reproduce the ensemble position distribution of several simple
systems. My presentation of this project will cover the theoretical
underpinnings and successful applications of our technique.
Part 2: Heuristic approaches to embedding NP Hard problems in Ising
models suitable for adiabatic quantum computation
A prominent architecture for prototype quantum annealing devices (in
particular the Dwave One) is designed to find the ground state of
Ising Hamiltonians with restricted couplings. Even if such a device
were to function in an idealized fashion there are many NP Hard
problems which cannot be mapped to this subset of Ising models at all,
which cannot be mapped in an efficient way, or for which finding the
mapping is itself an NP Hard problem. In the second part of my group
meeting I will discuss techniques that we've used which attempt to
find a heuristic embedding to such Hamiltonians. In particular, I will
discuss our work studying the optimization of water cluster
configurations with DWave One.
--
Ryan Babbush | PhD Student in Chemistry
(949) 331-3943 | babbush(a)fas.harvard.edu
Harvard University | Aspuru-Guzik Research Group
12 Oxford Street, Box 400 | Cambridge, MA 02138
Show replies by date