Applications of Extended Ensemble Monte Carlo

1
Applications of Extended Ensemble Monte Carlo

• Yukito IBA
• The Institute of Statistical Mathematics, Tokyo,
  Japan

2
(No Transcript)
3
Extended Ensemble MCMC

• A Generic Name which indicates
• Parallel Tempering,
• Simulated Tempering,
• Multicanonical Sampling,
• Wang-Landau,
• Umbrella Sampling

Valleau and Torrie
1970s
4
Contents

• 1. Basic Algorithms
• Parallel Tempering .vs Multicanonical
• 2. Exact Calculation with soft Constraints
• Lattice Protein / Counting Tables
• 3. Rare Events and Large Deviations
• Communication Channels
• Chaotic Dynamical Systems

5
Basic Algorithms

• Parallel Tempering
• Multicanonical Monte Carlo

6
References in physics

• Iba (2001) Extended Ensemble Monte Carlo
• Int. J. Mod. Phys. C12 p.623.
• A draft version will be found at
• http//arxiv.org/abs/cond-mat/0012323
• Landau and Binder (2005)
• A Guide to Monte Carlo Simulations in
  Statistical Physics (2nd ed. , Cambridge)
• A number of preprints will be found in
• Los Alamos Arxiv on the web.

This slide is added after the talk
7
Slow mixing by multimodal dist.
8
Bridging
fast mixing high temperature
slow mixing low temperature
9
Path Sampling
1.Facilitate Mixing 2.Calculate Normalizing
Constant (free energy)
Path Sampling Gelman and Meng (1998) stress
2. but 1. is also important
In Physics from 2. to 1. 1970s ? 1990s
10
Parallel Tempering

• a.k.a. Replica Exchange MC
• Metropolis Coupled MCMC

Geyer (1991), Kimura and Taki (1991) Hukushima
and Nemoto (1996) Iba(1993, in Japanese)
Simulate Many Replicas in Parallel
MCMC in a Product Space
11
Examples

• Gibbs Distributions with different temperatures
• Any Family parameterized by
• a hyperparameter

12
Exchange of Replicas
K4
13
Accept/Reject Exchange

• Calculate Metropolis Ratio
• Generate a Uniform Random Number
• in 0,1) and accept exchange
• iff

14
Detailed Balance in Extended Space
Combined Distribution
15
Multicanonical Monte Carlo
Berg et al. (1991,1992)

sufficient statistics
Energy not Expectation
Exponential Family
sufficient statistics
16
Density of States

The number of which satisfy
17
Multicanonical Sampling
18
Weight and Marginal Distribution
Original (Gibbs) Multicanonical
Random
19
flat marginal distribution
Scanning broad range of E
20
Reweighting

• Formally, for arbitrary it holds.
• Practically, is required,
  
• else the variance diverges in a large system.

21
Q. How can we do without knowledge on D(E)

• Ans.
• Estimate D(E) in the preliminary runs
• k th simulation

Simplest Method Entropic Sampling
in
22
Estimation of Density of States
(Ising Model on a random net)
30000 MCS
2
k1
3
5
4
10
14
11
k15
23
Estimation of D(E)

• Histogram
• Piecewise Linear
• Fitting, Kernel Density Estimation ..
• Wang-Landau
• Flat Histogram

Entropic Sampling
Original Multicanonical
Continuous Cases D(E)dE Non-trivial Task
24
Parallel Tempering / Multicanonical
parallel tempering combined distribution simula
ted tempering mixture distribution to
approximate
25
Potts model (2-dim, q10 states)
disordered
ordered
26
Phase Coexistence/ 1st order transition

• parameter (Inverse Temperature) changes
• sufficient statistics (Energy) jumps

water and ice coexists
27
Potts model (2-dim, q10 states)
disordered
ordered
28
Comparison

• _at_ Simple Liquids , Potts Models ..
• Multicanonical seems better than Parallel
  Tempering
• _at_ But, for more difficult cases ?
• ex. Ising Model with three spin Interaction

29
(No Transcript)
30
Soft Constraints

• Lattice Protein
• Counting Tables

The results on Lattice Protein are taken from
joint works with G Chikenji (Nagoya Univ) and
Macoto Kikuchi (Osaka Univ) Some examples are
also taken from the other works by Kikuchi and
coworkers.
31
Lattice Protein Model

• Motivation
• Simplest Models of Protein
• Lattice Protein
• Prototype of Protein-like molecules
• Ising Model
• Prototype of Magnets

32
Lattice Protein (2-dim HP)
33
sequence of
FIXED
and
corresponds to 2-types of amino acids (H and P)
conformation of chain
STOCHASTIC VARIABLE
SELF AVOIDING (SELF OVERLAP is not
allowed) IMPORTANT!
34
Energy (HP model)
the energy of conformation x is defined as
E(X) - the number of
in x
35
Examples
E -1
E0
Here we do not count the pairs neighboring on the
chain but it is not essential because the
difference is const.
36
MCMC

• Slow Mixing
• Even Non-Ergodicity with local moves

Bastolla et al. (1998) Proteins 32 pp. 52-66
Chikenji et al. (1999) Phys. Rev. Lett. 83
pp.1886-1889
37
Multicanonical

• Multicanonical w.r.t. E only
• NOT SUFFUCIENT
• Self-Avoiding condition is essential

38
Soft Constraint

• Self-Avoiding condition is essential

Soft Constraint
is the number of monomers that occupy the site i
39
Multi Self-Overlap Sampling

• Multi Self-Overlap Ensemble
• Bivariate Density of States
• in the (E,V) plane

V (self-overlap)
E
EXACT !!
40
Generation of Paths by softening of constraints
E
V0
large V
41
Comparison with multicanonical with hard
self-avoiding constraint
switching between three groups of minimum
energy states of a sequence
conventional (hard constraint)
proposed (soft constraint)
42
optimization
43
optimization (polymer pairs)
Nakanishi and Kikuchi (2006) J.Phys.Soc.Jpn. 75
pp.064803 / q-bio/0603024
44

double peaks
An Advantage of the method is that it can
use for the sampling at any temperature as well
as optimization
3-dim
Yue and Dill (1995) Proc. Nat. Acad. Sci. 92
pp.146-150
45
Another Sequence
Chikenji and Kikuchi (2000) Proc. Nat. Acad. Sci
97 pp.14273 - 14277
46
Related Works

• Self-Avoiding Walk without interaction /
  Univariate Extension
• Vorontsov-Velyaminov et al.
• J.Phys.Chem.,100,1153-1158 (1996)
• Lattice Protein but not exact / Soft-Constraint
  without control
• Shakhnovich et al.
• Physical Review Letters 67 1665 (1991)
• Continuous homopolymer -- Relax core
• Liu and Berne
• J Chem Phys 99 6071 (1993)
• See References in
• Extended Ensemble Monte Carlo, Int J Phys C
  12 623-656 (2001)
• but esp. for continuous cases,
• there seems more in these five years

47
(No Transcript)
48
Counting Tables
Pinn et al. (1998) Counting Magic Squares Soft
Constraints Parallel Tempering
49
Sampling by MCMC

• Multiple Maxima
• Parallel Tempering

50
Normalization Constant

calculated by Path sampling (thermodynamic
integration)
51
Latin square (3x3)
For each column, any given number appears once
and only once
For each raw, any given number appears once and
only once
52
Latin square (26x26)
This sample is taken from the web.
53
Counting Latin Squares

• 6
• 10
• 11

410000 MCS x 27 replicas
510000 MCS x 49 replicas
510000 MCS x 49 replicas
other 3 trials
54
Counting Tables

• Soft Constraints Extended Ensemble MC
• Quick and Dirty ways of calculating the number
  of tables that satisfy given constraints.
• It may not be optimal for a special case,
• but no case-by-case tricks, no mathematics,
• and no brain is
  required.

55
(No Transcript)
56
Rare Events and Large Deviations

• Communication Channels 1
• Chaotic Dynamical Systems 2

1 Part of joint works with Koji
Hukushima (Tokyo Univ).
2 Part of joint works with Tatsuo
Yanagita (Hokkaido Univ). (The result shown
here is mostly due to him )
57
Applications of MCMC

• Statistical Physics (1953 )
• Statistical Inference (1970s,1980s, 1990)
• Solution to any problem on
• sampling counting
• estimation of large deviation
• generation of rare events

58
Noisy Communication Channel
prior
encoded degraded
decode
distance (bit errors)
by Viterbi, loopy BP, MCMC
59
Distribution of Bit Errors
Kronecker delta
tails of the distribution is not easy to estimate
60
Introduction of MCMC
NOT sampling from the posterior

• Sampling noise in channels by the MCMC
• Given an error-correcting code
• Some patterns of noise are very harmful
• difficult to correct
• Some patterns of noise are safe
• easy to correct

61
Multicanonical Strategy

• MCMC sampling of
• Broad distribution of
• ? Broad distribution of distance

and
62
Multicanonical Sampling

• MCMC Sampling and
• with the weight

Estimated by the iteration of preliminary runs
exactly what we want, but can be ..
63
flat marginal distribution
Enable efficient calculation of the tails of the
distribution (large deviation)
Scanning broad range of bit errors
64
Example

• Convolutional Code

Viterbi decoding
Binary Symmetric Channel Fix the number of
noise (flipped bits)
65
Simplification

• In this case

is independent of
Set
Binary Symmetric Channel Fix the number of
noise (flipped bits)
sum over the possible positions of the noise
66
Simulation
difficult to calculate by simple sampling
the number of bit errors
67
Correlated Channels

• It will be useful for the study of
  error-correcting code in a correlated channel.
• Without assuming models of correlation
• in the channel we can sample relevant
• correlation patterns.

68
(No Transcript)
69
Rare events in Dynamical Systems

• Deterministic Chaos
• Doll et al. (1994), Kurchan et al. (2005)
• Sasa, Hayashi, Kawasaki .. (2005 )
• (Mostly) Stochastic Dynamics
• Chandler Group
• Frenkel et al.
• and more

Stagger and Step Method Sweet, Nusse, and Yorke
(2001)
Transition Path Sampling
70
Sampling Initial Condition

• Sampling initial condition of
• Chaotic dynamical systems
• Rare Events

71
Double Pendulum
72
Unstable fixed points
control and stop the pendulum one of the three
positions
energy dissipation (friction) is assumed i.e.,
no time reversal sym.
73
Definition of artificial energy
stop zero velocity
stopping position
penalty to long time
T is max time
74
Metropolis step
Evaluate Energy
Perturb Initial State
Integrate Equation of Motion and Simulate
Trajectory
Reject or Accept
75
? Parallel Tempering
76
An animation by Yanagita is shown in the talk,
but might not be seen on the web.
77
Summary

• Extended Ensemble Soft Constraint strategy
  gives simple solutions to a number of difficult
  problems
• The use of MCMC should not be restricted to the
  standard ones in Physics and Bayesian Statistics.
  
• To explore new applications of MCMC extended
  ensemble MC will play an essential role.
  

78
END