NIMBLE is a modeling language and algorithm programming framework for Bayesian and likelihood-based statistical analysis. It allows users to write custom algorithms that operate on statistical models specified in the BUGS language. The framework processes BUGS models to extract relationships, builds a graphical model object, generates C++ code, and provides interfaces to compiled algorithm functions. This allows for flexible and distributed development of advanced Bayesian computational methods.

1. NIMBLE
(First
release
by
summer)
model
language
(BUGS)
Algorithm
language

2. Why
algorithm
programmability?
• Simple
advanced
MCMC:
block
updaters,
mulHple
updaters,
mulHple
scales,
modularity
• Advanced
advanced
MCMC:
bridge
sampling,
reversible
jump,
adapHve
MCMC,
tempering,
approximaHng
models
• ParHcle
filtering
• Importance
sampling
• Laplace
approximaHon,
adapHve
Gaussian
quadrature
• Kalman
Filtering,
Unscented
KF,
extended
KF
• Prior
sensiHvity
analysis
by
reweighHng
• Monte
Carlo
ExpectaHon
MaximizaHon
(MCEM)
• Data
cloning
• Approximate
Bayesian
ComputaHon
• Posterior
predicHve
checks
(without
re-­‐running
MCMC)
• Normalizing
constants:
bridge
sampling;
see
above
• CombinaHon
algorithms:
PF
+
MCMC,
MCMC
+
Laplace
Algorithm
developers
need
a
system
for
model-­‐flexibility
and
distribuHon.

3. What
we
do
with
BUGS
models
model
{
mu
~
dnorm(0,
10)
logit(p)
~
dnorm(mu,
sigma)
y
~
dbin(N,
p)
}

4. What
we
do
with
BUGS
models
model
{
mu
~
dnorm(0,
10)
logit(p)
~
dnorm(mu,
sigma)
y
~
dbin(N,
p)
}
Extract
all
semanHc
relaHonships.
Build
a
graphical
model
object
R
(igraph).
Generate
C++
code
Compile,
load,
and
provide
interface
objects.

5. What
we
do
with
BUGS
models
model
{
mu
~
dnorm(0,
10)
logit(p)
~
dnorm(mu,
sigma)
y
~
dbin(N,
p)
}
Extract
all
semanHc
relaHonships.
Build
a
graphical
model
object
R
(igraph).
Generate
C++
code
Compile,
load,
and
provide
interface
objects.
mymodel$mu
<-­‐
5
toCalc
<-­‐
getDependencies(mymodel,
"mu")
calculate(mymodel,
toCalc)

6. What
we
do
with
BUGS
models
Some
"automaHc"
extensions
of
BUGS
1. Expressions
as
arguments
2. MulHple
parameterizaHons
for
distribuHons
(e.g.
precision
or
std.
deviaHon)
3. Compile-­‐Hme
if-­‐then-­‐else
4. Single-­‐line
moHfs:
y[1:10]
~
glmm(X
*
A
+
(1
|
Group),
family
=
"binomial")
model
{
mu
~
dnorm(0,
10)
logit(p)
~
dnorm(mu,
sigma)
y
~
dbin(N,
p)
}
Extract
all
semanHc
relaHonships.
Build
a
graphical
model
object
R
(igraph).
Generate
C++
code
Compile,
load,
and
provide
interface
objects.
mymodel$mu
<-­‐
5
toCalc
<-­‐
getDependencies(mymodel,
"mu")
calculate(mymodel,
toCalc)

7. How
we
write
algorithms
Random-­‐walk
MCMC
updater
1.
setup
arguments:
model,
targetNode

8. How
we
write
algorithms
Random-­‐walk
MCMC
updater
1.
setup
arguments:
model,
targetNode
2.
setup
code:
use
model
structure
to
query
targetNode
dependencies
Processed
ONCE,
in
R

9. How
we
write
algorithms
Random-­‐walk
MCMC
updater
1.
setup
arguments:
model,
targetNode
2.
setup
code:
use
model
structure
to
query
targetNode
dependencies
Processed
ONCE,
in
R
3.
run-­‐Hme
code:
(i)
put
a
proposal
value
in
the
targetNode
(ii)
calculate
needed
log
likelihoods
(iii)
accept
or
reject
Becomes
C++
code,
compiled
&
interfaced

10. How
we
write
algorithms
Random-­‐walk
MCMC
updater
1.
setup
arguments:
model,
targetNode
2.
setup
code:
use
model
structure
to
query
targetNode
dependencies
Processed
ONCE,
in
R
3.
run-­‐Hme
code:
(i)
put
a
proposal
value
in
the
targetNode
(ii)
calculate
needed
log
likelihoods
(iii)
accept
or
reject
Becomes
C++
code,
compiled
&
interfaced
Building
a
program
with
many
funcHons:
• model
structure
(setup
arguments)
processed
in
R
• algorithm
executed
in
C++

11. How
we
write
algorithms
Random-­‐walk
MCMC
updater
1.
setup
arguments:
model,
targetNode
2.
setup
code:
use
model
structure
to
query
targetNode
dependencies
Processed
ONCE,
in
R
3.
run-­‐Hme
code:
(i)
put
a
proposal
value
in
the
targetNode
(ii)
calculate
needed
log
likelihoods
(iii)
accept
or
reject
Becomes
C++
code,
compiled
&
interfaced
Building
a
program
with
many
funcHons:
• model
structure
(setup
arguments)
processed
in
R
• algorithm
executed
in
C++
We
provide
data
structures
for
sets
of
model
values
(e.g.
MCMC
output)

12. ConnecHons
to
other
engines
When
another
system
has
a
great
algorithm,
we
want
to
wrap
access
to
it.

13. ConnecHons
to
other
engines
When
another
system
has
a
great
algorithm,
we
want
to
wrap
access
to
it.
Example:
Monte
Carlo
ExpectaHon
MaximizaHon
(MCEM)
• Uses
MCMC
repeatedly
on
latent
states
• Parameters
updated
between
MCMCs
• We
don't
have
to
use
our
own
MCMC
if
someone
else
has
a
great
one

14. ConnecHons
to
other
engines
When
another
system
has
a
great
algorithm,
we
want
to
wrap
access
to
it.
Example:
Monte
Carlo
ExpectaHon
MaximizaHon
(MCEM)
• Uses
MCMC
repeatedly
on
latent
states
• Parameters
updated
between
MCMCs
• We
don't
have
to
use
our
own
MCMC
if
someone
else
has
a
great
one
How?
We
have
high-­‐level
representaHon
of
model
structure.
Can
generate
specificaHons
needed
for
other
packages.

15. Who
we
think
will
use
NIMBLE
1.
Algorithm
developers
2.
Data
analysts
3.
Workflow
pipelines

16. Numerical!
Integration of!
Mixture Models for!
Bayesian and !
Likelihood!
Estimation!
Team:!
Perry de Valpine!
Chris Paciorek!
Daniel Turek!
Cliff Anderson-Bergmann!
Ras Bodik!
Duncan Temple Lang!
Funding:!
NSF Advances in !
Biological Informatics!

18. Common
situaHon
model
language
(Black
box
algorithm)

19. How
we
build
an
MCMC
1.
R
funcHons
use
model/graph
objects
in
R
to
inspect
a
parHcular
model:
-­‐
Build
list
of
updaters
-­‐
Can
be
inspected,
modified
-­‐
You
can
write
your
own

20. How
we
build
an
MCMC
1.
R
funcHons
use
model/graph
objects
in
R
to
inspect
a
parHcular
model:
-­‐
Build
list
of
updaters
-­‐
Can
be
inspected,
modified
-­‐
You
can
write
your
own
2.
Generate
C++
code
for
all
updaters.

21. How
we
build
an
MCMC
1.
R
funcHons
use
model/graph
objects
in
R
to
inspect
a
parHcular
model:
-­‐
Build
list
of
updaters
-­‐
Can
be
inspected,
modified
-­‐
You
can
write
your
own
2.
Generate
C++
code
for
all
updaters.
3.
Compile,
load,
build
and
interface
to
objects.