Cameron T Chalk Bin Fu Alejandro Huerta Mario A Maldonado Eric Martinez Robert T Schweller Tim Wylie Funding by NSF Grant CCF1117672 NSF Early Career Award 0845376 Introduction ID: 776448
Download Presentation The PPT/PDF document " Flipping Tiles: Concentration Independe..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Flipping Tiles: Concentration Independent Coin Flips in Tile Self-Assembly
Cameron T. Chalk, Bin Fu, Alejandro Huerta, Mario A. Maldonado, Eric Martinez, Robert T. Schweller, Tim WylieFunding by NSF Grant CCF-1117672NSF Early Career Award 0845376
?
?
Slide2Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide3?
?
Slide4Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide5Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
Slide6Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide7Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide8Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide9Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide10Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide11Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide12Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide13Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide14Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide15Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide16Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide17Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
Slide18Tile Assembly Model(Rothemund, Winfree, Adleman)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
y
) = 2
G(
r
) = 2
G(
b
) = 1
G(
p
) = 1
S
Temperature: 2
Seed:
S
TERMINAL
Slide19Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
p
) = 2
G(b) = 2
S
Temperature: 2
Seed:
.1
.2
.2
.2
.3
Slide20Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
p
) = 2
G(b) = 2
S
Temperature: 2
Seed:
S
.1
.2
.2
.2
.3
Slide21Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
p
) = 2
G(b) = 2
S
Temperature: 2
Seed:
S
S
S
.1
.2
.2
.2
.3
Slide22Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
p
) = 2
G(b) = 2
S
Temperature: 2
Seed:
.1
.2
.2
.2
.3
S
S
.2
.2 + .3
S
=
.4
Slide23Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
Tileset
:
S
Glue:
G(
g
) = 2
G(
o
) = 2
G(
p
) = 2
G(b) = 2
S
Temperature: 2
Seed:
.1
.2
.2
.2
.3
S
S
.2
.2 + .3
S
=
.4
.3
.2 + .3
=
.6
Slide24Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
Slide25Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
Slide26Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
S
S
1
.5
.5
1
Slide27Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
S
S
1
.5
.5
1
(.4)(.5)(1)
Slide28Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
S
S
1
.5
.5
1
(.4)(.5)(1)
+
(.4)(.5)(.5)
Slide29Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
S
S
1
.5
.5
1
(.4)(.5)(1)
+ (.4)(.5)(.5)
+
(.6)(.5)(.5)
.45
Slide30Probabilistic Tile Assembly Model(Becker, Remila, Rapaport)
S
.1
.2
.2
.2
.3
.4
.6
S
S
S
S
S
S
.5
.5
.5
.5
S
S
1
.5
.5
1
+ (.6)(.5)(.5)
+ (.6)(.5)(1)
.55
(.4)(.5)(1)
+ (.4)(.5)(.5)
+ (.6)(.5)(.5)
.45
(.4)(.5)(.5)
Slide31Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide32Concentration Independent Coin Flipping
(TAS, C)
Slide33Concentration Independent Coin Flipping
(TAS, C) { , , , , }
Slide34Concentration Independent Coin Flipping
(TAS, C) { , , , , }
Slide35Concentration Independent Coin Flipping
(TAS, C) { , , , , }
P( ) + P( )
+
P(
) = .5
Slide36Concentration Independent Coin Flipping
(TAS, C) { , , , , }
P( ) + P( )
+
P(
) = .5
P( ) + P(
) = .5
Slide37Concentration Independent Coin Flipping
(TAS, C) { , , , , }
P( ) + P( )
+
P(
) = .5
P( ) + P(
) = .5
For ALL C
Slide38.5
.5
Slide39.5
.5
.5
.5
Slide40.5
.5
.5
.5
P(
) = .5
P(
) = .5
Slide41.7
.3
.3
.7
Slide42.7
.3
.3
.7
P(
) = .3
P(
) = .7
Slide43Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide44x
y
Slide45x
y
Slide46x
y
Slide47x
y
Slide48x
y
Slide49x
y
x
x
+y
y
y+y
1
P( ) =
x
x
+y
y
y+y
Slide50x
y
x
x
+y
y
y+y
1
P( ) =
xy
2y(
x+y
)
Slide51x
y
x
x
+y
P( ) =
xy
y
y+y
y
x
+y
+
x
x
+y
y
y+y
y
x
+y
2y(
x+y
)
Slide52x
y
x
x
+y
P( ) =
xy
y
y+y
y
x
+y
+
xy
2
2y(
x+y
)
2
2y(
x+y
)
Slide53x
y
P( ) =
xy
y
x
+y
+
xy
2
2y(
x+y
)
2
x
x+x
y
x
+y
2y(
x+y
)
Slide54x
y
P( ) =
xy
y
x
+y
+
xy
2
2y(
x+y
)
2
x
x+x
y
x
+y
y
x+y
+
x
x+x
y
x
+y
2y(
x+y
)
Slide55x
y
P( ) =
xy
y
x
+y
+
xy
2
2y(
x+y
)
2
x
x+x
y
x
+y
xy
2
+
2x(
x+y
)
2
2y(
x+y
)
Slide56P( ) =
xy
+
2y(
x+y
)
2
xy
2
+
2x(
x+y
)
2
2y(
x+y
)
x
2
+ 2xy + y
2
2(
x+y
)
2
=
=
(
x+y
)
2
2(
x+y
)
2
=
1
2
xy
2
Slide57Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide58S
1
Slide59S
1
1
2
Slide60S
1
1
2
2
3
Slide61S
1
1
2
2
3
3
4
Slide62S
1
1
2
2
3
3
4
4
5
Slide63S
1
1
2
2
3
3
4
4
5
5
6
Slide64S
1
1
2
2
3
3
4
4
5
5
6
6
7
Slide65S
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
Slide66S
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
Slide67S
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
Slide68S
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
Slide69S
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
Slide70S
Slide71S
Slide72S
S
S
Slide73S
S
S
S
S
S
Slide74S
S
S
S
S
S
S
S
Slide75Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide76Simulation
(TAS, C) { , , , , }
Slide77Simulation
(TAS, C) { , , , , }
(
TAS’, for all C
) { , , , , }
with m x n scale factor
Slide78Simulation
(TAS, C) { , , , , }
(
TAS’, for all C
) { , , , , }
with m x n scale factor
P(TAS, C) = P(TAS’, for all C)
Slide79Simulation
(TAS, C) { , , , , }
(
TAS’, for all C
) { , , , , }
with m x n scale factor
P(TAS, C) = P(TAS’, for all C)
TAS’ robustly simulates TAS for C at scale factor
m,n
Simulation
Unidirectional two-choice linear assembly systems:
Grows in one direction from seed
Non-determinism between at most 2 tiles
Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
Slide82Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
a
b
S
S
Slide83Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
a
b
S
S
a
S
Slide84Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
a
b
S
S
a
S
b
S
c
b
S
d
Slide85Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
a
b
S
S
a
S
b
S
c
b
S
d
b
S
d
Slide86Simulation
Unidirectional two-choice linear assembly systems:Grows in one direction from seedNon-determinism between at most 2 tiles
S
a
b
S
S
a
S
b
S
c
b
S
d
b
S
d
For any unidirectional two-choice linear assembly system X, there exists a tile assembly system X’ which robustly simulates X for the uniform concentration distribution at scale factor 5,4.
Slide87Simulation
S
S
S
Slide88Simulation
S
S
S
S
Slide89Simulation
S
S
S
S
S
S
Slide90Simulation
S
S
S
S
S
S
Slide91Simulation
S
S
S
S
S
S
Slide92Simulation
S
S
S
S
S
S
Slide93Simulation
S
S
S
S
S
S
Slide94Simulation
S
S
S
S
S
S
Slide95Simulation
S
S
S
S
S
S
Slide96Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide97Simulation Application
There exists a TAS which assembles an expected length
N
linear assembly using
Θ
(
log
N
) tile types.
(
Chandran
,
Gopalkrishnan
,
Reif
)
Slide98Simulation Application
There exists a TAS which assembles an expected length
N
linear assembly using
Θ
(
log
N
) tile types.
(
Chandran
,
Gopalkrishnan
,
Reif
)
The construction is a unidirectional two-choice linear assembly system
Applies to uniform concentration distribution
Slide99Simulation Application
There exists a TAS which assembles an expected length
N
linear assembly using
Θ
(
log
N
) tile types.
(
Chandran
,
Gopalkrishnan
,
Reif
)
The construction is a unidirectional two-choice linear assembly system
Applies to uniform concentration distribution
Corollary to simulation technique:
There exists a TAS which assembles a width-4 expected length
N
assembly for all concentration distributions using O(
log
N
) tile types.
Slide100Simulation Application
There exists a TAS which assembles an expected length
N
linear assembly using
Θ
(
log
N
) tile types.
(
Chandran
,
Gopalkrishnan
,
Reif
)
The construction is a unidirectional two-choice linear assembly system
Applies to uniform concentration distribution
Corollary to simulation technique:
There exists a TAS which assembles a width-4 expected length
N
assembly for all concentration distributions using O(
log
N
) tile types.
Further, there is no PTAM tile system which generates width-1 expected length
N
assemblies for all concentration distributions with less than
N
tile types.
Slide101Introduction
Models
Concentration Independent Coin Flip
Big Seed, Temperature 1
Single Seed, Temperature 2
Simulation
Simulation Application
Unstable Concentrations
Summary
Slide102Unstable Concentrations
(TAS, C) { , , , , }
P( ) + P( )
+
P(
) = .5
P( ) + P(
) = .5
Slide103Unstable Concentrations
(TAS, C) { , , , , }
P( ) + P( )
+
P(
) = .5
P( ) + P(
) = .5
At each assembly stage, C changes
Slide104Unstable Concentrations
Impossible in the
aTAM
(in bounded space
)
Slide105Unstable ConcentrationsImpossible in the aTAM (in bounded space)Possible (and easy) in some extended models:
aTAM
with neg.
Interactions, big seed
Slide106Unstable ConcentrationsImpossible in the aTAM (in bounded space)Possible (and easy) in some extended models:
aTAM
with neg. Interactions, big seed
Hexagonal TAM with neg.
Interactions
Slide107Unstable ConcentrationsImpossible in the aTAM (in bounded space)Possible (and easy) in some extended models:
aTAM
with neg. Interactions, big seed
Hexagonal TAM with neg.Interactions
Polyomino
TAM
Slide108Unstable ConcentrationsImpossible in the aTAM (in bounded space)Possible (and easy) in some extended models:
aTAM
with neg. Interactions, big seed
Hexagonal TAM with neg.Interactions
Polyomino TAM
Geometric TAM,
big seed
Slide109Summary
Unstable Concentrations
Impossible in the aTAM (in bounded space)Extended models:Future Work:Single seed temperature 1Other simulations (other TASs/Boolean circuits)Uniform random number generationRandomized algorithms
Simulation
Concentration Independent Coin FlipsLarge seed, temperature 1Single seed, temperature 2
S
S
S
S
S
S
S