20140419 kawajiri m1gp
TRANSCRIPT
-
Online Distributed Sensor Selection,Daniel Golovin, Matthew Faulkner, Andreas Krause,
IPSN2010
2014/04/19 M1GP 1
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
2
4
(Sensor Selection)
2
2
etc.
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
1. etc.
2. ()
3.
3
? ?
???
? ?
???
? ?
???
? ?
???
? ?
???
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
V , |V| = N k
4
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
t
5
2 3 2
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
t
6
[Streeter & Golovin 08]: Online Greedy (OG)
Value of
-
Intelligent Cooperative Systems Lab. The University of TokyoICS 7
[Auer et al 95]: EXP3
N
EXP3 [Auer et al 95]
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
8
P(1) P(2) P(3)
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
9
: EXP3 4 P(1) P(2) P(3)
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
4610 : Intel Research Berkeley
(EMSE)
10
SERVER
LAB
KITCHEN
COPYELEC
PHONEQUIET
STORAGE
CONFERENCE
OFFICEOFFICE50
51
52 53
54
46
48
49
47
43
45
44
42 41
3739
38 36
33
3
6
10
11
12
13 14
1516
17
19
2021
22
242526283032
31
2729
23
18
9
5
8
7
4
34
1
2
3540
()
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
()
(10/46)
11
Oine greedy
Distributed Online Greedy
96% (9.48 / 9.85)
99% (9.74 / 9.85)
(
)
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
1
2
12
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
() A. Krause : NIPS, ICML, IPSN, IROS, ICRA, CVPR,
13
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
14
-
Intelligent Cooperative Systems Lab. The University of TokyoICS
Tutorial Andreas Krause http://submodularity.org/
15
-
Online Distributed Sensor SelecRon
Daniel Golovin, MaUhew Faulkner, Andreas Krause
rsrg @caltech ..where theory and practice collide 16
-
Sensor-equipped cell phones are ubiquitous.
Which sensors should send data?
Can current measurements inform selecHon?
17
Community Sensing
Used for trac monitoring, polluRon detecRon, earthquake measurement.
Constraints on bandwidth, power, privacy ImpracRcal to query all phones.
-
4
Select two cameras to query, in order to detect the most people.
18
A Sensor SelecRon Problem
People Detected:
2
Duplicates only counted once
-
Set V of sensors, |V| = N Select a set of k sensors Sensing quality model
Typically NP-hard
A Sensor SelecRon Problem
19
-
20
Submodularity Diminishing returns property for adding more sensors.
Many objec0ves are submodular: DetecRon, coverage, mutual informaRon, and others.
+2
+1
For all , and a sensor ,
-
Lets choose sensors S = {v1 , , vk} greedily
[Nemhauser et al 78] If F is submodular, the Greedy algorithm gives constant factor approximaRon:
Greedy SelecRon
1. Must know sensing model F 2. Greedy is centralized 3. SelecHon ignores current
sensor values 21
-
22
Online Sensor SelecRon Get to choose sensors on each round t. Then is revealed.
Need to explore dierent sets.
Only need to evaluate F for chosen sets.
2 3 2
-
23
Online Sensor SelecRon Get to choose sensors on each round t. Then is revealed.
Round 1 Round 2 Round 3
Only assume is submodular and bounded
-
24
Online Greedy SelecRon At each round, choose a set . Learn to choose greedily.
Theorem [Streeter & Golovin 08]: Online Greedy (OG) The centralized Online Greedy algorithm chooses
Value of What algorithm?
-
25
On each round, choose one sensor and observe it value.
Theorem [Auer et al 95]: The average value obtained by EXP3 converges to the value of the xed opRmum:
Single Sensor SelecRon
EXP3 [Auer et al 95]
balances exploring and exploiRng
Can we avoid centralized sampling?
-
26
Idea: Independent draws unRl exactly one sensor broadcasts a success.
Distributed Sampling
Doesnt sample from correct distribuHon
P(1) P(2) P(3)
Centralized sampling may not scale pracRcally.
-
27
A Distributed Sampling Protocol
Theorem: Protocol correctly samples from P. Requires < 4 messages in the broadcast model
We can sample from correct distribuRon, while using few messages!
P(1) P(2) P(3)
-
28
Use distributed sampling protocol in EXP3. Yields distributed single-sensor selecRon algorithm
Distributed EXP3
Broadcast the change of weight for now
Distributed EXP3
Theorem: Exact same performance as centralized EXP3
-
29
Distributed Online Greedy Distributed Online Greedy (DOG) selects a set of k sensors on each round, using Distributed EXP3 as a subrouRne.
D-EXP3 D-EXP3 D-EXP3
Theorem : DOG selects sensors St that obtain
Using messages per round in expectaRon.
-
30
SelecRon techniques extend eciently to non-broadcast communicaRon models.
CommunicaRon Models
Star Network Model: messages between base staRon and one sensor are unit cost.
D-EXP3 samples from Each sensor needs to know the sum of all
weights
Lazy-DOG. A sensor only updates its sum when it communicates with base staRon.
Theorem: Lazy-DOG gives same selecRon performance as DOG, and reduces messages in star model from N to log(N).
-
31
ObservaRon-Dependent SelecRon Sensing can be cheap while communicaRon is costly. Can current observaRons inform selecRon?
Valuable observaHon Domain
knowledge
-
32
ObservaRon-Dependent SelecRon
2. Sensor v acRvates if exceeds a threshold.
3. Given communicaRon cost C, feed back
OD-DOG. A sensors current measurement can inuence its decision to acRvate.
1. Each sensor v esRmates its marginal value
Learn the threshold
Useful for detecHng important and rare events
-
33
Temperature Monitoring Select 10 from 46 temperature sensors deployed at Intel Research Berkeley.
SERVER
LAB
KITCHEN
COPYELEC
PHONEQUIET
STORAGE
CONFERENCE
OFFICEOFFICE50
51
52 53
54
46
48
49
47
43
45
44
42 41
3739
38 36
33
3
6
10
11
12
13 14
1516
17
19
2021
22
242526283032
31
2729
23
18
9
5
8
7
4
34
1
2
3540
OpRmize the expected reducRon in mean squared predicRon error (EMSE).
(oten) submodular*
-
34
Temperature Monitoring
Oine greedy
Distributed Online Greedy
OpRmize sensor placement for monitoring temperature in an oce building. Select 10 of 46 sensors.
-
35
Outbreak DetecRon Ba
-
36
Outbreak DetecRon
High communicaHon
cost
Low communicaHon cost
Balances added value and communicaHon cost
Greedy
0.1 avg. extra acHvaHons
5 avg. extra acHvaHons
OD-DOG with observaRon-dependent selecRon for various communicaRon costs C.
-
DOG, a distributed sensor selecRon algorithm that applies to many sensing applicaRons.
Strong theoreRcal guarantees on performance and communicaRon cost.
OD-DOG for observaRon-specic selecRon. Can incorporate domain knowledge.
Performs well on several real sensor data sets.
Conclusions
37