change detection in dynamic environments mark steyvers scott brown uc irvine this work is supported...

36
Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific Research (AFOSR grant number FA9550-04-1-0317)

Post on 19-Dec-2015

213 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Change Detection in Dynamic Environments

Mark Steyvers

Scott Brown

UC Irvine

This work is supported by a grant from the US Air Force Office of Scientific Research (AFOSR grant number FA9550-04-1-0317)

Page 2: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Overview

• Experiments with dynamically changing environments

• Task: Given a sequence of random numbers, predict the next one

• Questions:

– How do observers detect changes?

– What are the individual differences? “Jumpiness”

• Bayesian models + simple process models

Page 3: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

= observed data

= prediction

Two-dimensional prediction task

• 11 x 11 button grid• Touch screen monitor• 1500 trials • Self-paced • Same sequence for all

subjects

Page 4: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0

5

10

0

5

10

Subject 4

40 50 60 70 80 90 100 110 120 1300

5

10

Time

Subject 12

Sequence Generation

• (x,y) locations are drawn from two binomial distributions of size 10, and parameters θ

• At every time step, probability 0.1 of changing θ to a new random value in [0,1]

• Example sequence:

Time

θ=.12 θ=.95 θ=.46 θ=.42 θ=.92 θ=.36

Page 5: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Example Sequence

Page 6: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0

5

10

Optimal Bayesian Solution

0

5

10

Subject 4

40 50 60 70 80 90 100 110 120 1300

5

10

Time

Subject 12

= observed sequence

Bayesian Solution= prediction

0

5

10

Optimal Bayesian Solution

0

5

10

Subject 4

40 50 60 70 80 90 100 110 120 1300

5

10

Time

Subject 12

Subject 4 – change detection too slow0

5

10

Optimal Bayesian Solution

0

5

10

Subject 4

40 50 60 70 80 90 100 110 120 1300

5

10

Time

Subject 12

Subject 12 – change detection too fast

(sequence from block 5)

Page 7: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Tradeoffs

• Detecting the change too slowly will result in lower accuracy and less variability in predictions than an optimal observer.

• Detecting the change too quickly will result in false detections, leading to lower accuracy and higher variability in predictions.

Page 8: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0.5 1 1.5 2 2.5 3 3.52.7

2.8

2.9

3

3.1

3.2

3.3

3.4

3.5

Mean Absolute Movement

Me

an

Abs

olu

te T

ask

Err

or

12

34

5

6

7

8 9

OPTIMAL SOLUTION

Average Error vs. Movement

= subject

Page 9: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

“Ideal” observer: inferring the HMM that generated the data

2x

2

2y

1x

1

1y

tx

t

ty

...

Time 1 2 t t+1

Measurements

states

changepoints

Change probability

1tx

1t

?

Page 10: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Gibbs sampling

Model predictionInferred changepoint

• Sample from distribution over change points. • Prediction is based on average measurement after last inferred

changepoint

Page 11: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0.5 1 1.5 2 2.5 3 3.52.7

2.8

2.9

3

3.1

3.2

3.3

3.4

3.5

Mean Absolute Movement

Me

an

Abs

olu

te T

ask

Err

or

12

34

5

6

7

8 9

OPTIMAL SOLUTION

Average Error vs. Movement

= subject

Page 12: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

A simple process model

1. Make new prediction some fraction α of the way between recent outcome and old prediction α = change proportion

2. Fraction α is a linear function of the error made on last trial

3. Two free parameters: A, B

A<B bigger jumps with higher error

A=B constant smoothing

1 (1 )t t tp y p

t t tError y p

α

0

1

A

B

BA

Page 13: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Average Error vs. Movement

0.5 1 1.5 2 2.5 3 3.52.7

2.8

2.9

3

3.1

3.2

3.3

3.4

3.5

Mean Absolute Movement

Me

an

Abs

olu

te T

ask

Err

or

12

34

5

6

7

8 9

OPTIMAL SOLUTION

= subject

= model

Page 14: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

One-dimensional Prediction Task

1

2

3

4

5

6

7

8

9

10

11

12

12 PossibleLocations

• Where will next blue square arrive on right side?

Page 15: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Average Error vs. Movement

0 0.5 1 1.5 2

1.2

1.4

1.6

1.8

2

2.2

2.4

Mean Absolute Movement

Me

an

Abs

olu

te T

ask

Err

or

12

3

4

56

7

8

910

11

12

13

14

15

16

17

1819

2021

OPTIMAL SOLUTION

= subject

= model

Page 16: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

New Experiments

• Prediction judgments might not be best measurement for assessing psychological change

• New experiments:

– Inference judgment: what currently is the state of the system?

Page 17: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Inference Task(aka filtering)

2x

2

2y

1x

1

1y

tx

t

ty

...

Time 1 2 t t+1

1tx

?

1ty

What is the cause of yt+1?

Page 18: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Tomato Cans Experiment

• Cans roll out of pipes A, B, C, or D

• Machine perturbs position of cans (normal noise)

• At every trial, with probability 0.1, change to a new pipe (uniformly chosen)

(real experiment has response buttons and is subject paced)

A B C D

Page 19: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Tomato Cans Experiment

(real experiment has response buttons and is subject paced)

A B C D • Cans roll out of pipes A, B, C, or D

• Machine perturbs position of cans (normal noise)

• At every trial, with probability 0.1, change to a new pipe (uniformly chosen)

• Curtain obscures sequence of pipes

Page 20: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Tasks

A B C D• Inference:

what pipe produced the last can?

A, B, C, or D?

Page 21: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Cans Experiment

• 136 subjects

• 16 blocks of 50 trials

• Vary change probability across blocks

– 0.08

– 0.16

– 0.32

• Question: are subjects sensitive to the number of changes?

Page 22: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

50

60

70

80

90

100low alpha

50

60

70

80

90

100

% A

ccu

racy

(ag

ain

st T

rue

)med alpha

0 10 20 30 40 50 6050

60

70

80

90

100

% Changes

high alpha

ideal

ideal

ideal

Prob. = .08

Prob. = .16

Prob. = .32

Page 23: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Plinko/ Quincunx Experiment

Physical version Web version

Page 24: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Conclusion

• Adaptation in non-stationary decisionenvironments

• Individual differences

– Over-reaction: perceiving too much change

– Under-reaction: perceiving too little change

Page 25: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Do the experiments yourself:

http://psiexp.ss.uci.edu

Page 26: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Number of Perceived Changes per Subject

Low medium high

Change Probability

(Red line shows ideal number of changes)

Subject #1

Page 27: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Number of Perceived Changes per Subject

55% of subjects show increasing pattern

45% of subjects show non-increasing pattern

Low, medium, high change probability Red line shows ideal number of changes

Page 28: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Tasks

A B C D• Inference:

what pipe produced the last can?

A, B, C, or D?

• Prediction: in what region will the next can arrive?

1, 2, 3, or 4?

1 2 3 4

Page 29: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

Cans Experiment 2

• 63 subjects

• 12 blocks

– 6 blocks of 50 trials for inference task

– 6 blocks of 50 trials for prediction task

– Identical trials for inference and prediction

Page 30: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

INFERENCE PREDICTION

Sequence

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

1

2

3

4

5

6

1 2 3 4 5 6A B C D

Trial

Page 31: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

INFERENCE PREDICTION

Sequence

Ideal Observer

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

1

2

3

4

5

6

1 2 3 4 5 6A B C D

Trial

Page 32: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

INFERENCE PREDICTION

Sequence

Ideal Observer

Individualsubjects

Trial0 0.2 0.4 0.6 0.8 1

0

0.2

0.4

0.6

0.8

1

1

2

3

4

5

6

1 2 3 4 5 6A B C D

Page 33: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

INFERENCE PREDICTION

Sequence

Ideal Observer

Trial0 0.2 0.4 0.6 0.8 1

0

0.2

0.4

0.6

0.8

1

1

2

3

4

5

6

1 2 3 4 5 6A B C D

Individualsubjects

Page 34: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

INFERENCE PREDICTION

Sequence

Ideal Observer

Trial0 0.2 0.4 0.6 0.8 1

0

0.2

0.4

0.6

0.8

1

1

2

3

4

5

6

1 2 3 4 5 6A B C D

Individualsubjects

Page 35: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0 20 40 6030

40

50

60

70

80

90

100

Changes (%)

Acc

ura

cy

0 20 40 6030

40

50

60

70

80

90

100

Changes (%)

Acc

ura

cy

= Subjectideal

ideal

INFERENCE PREDICTION

Page 36: Change Detection in Dynamic Environments Mark Steyvers Scott Brown UC Irvine This work is supported by a grant from the US Air Force Office of Scientific

0 20 40 6030

40

50

60

70

80

90

100

Changes (%)

Acc

ura

cy

ideal

0 20 40 6030

40

50

60

70

80

90

100

Changes (%)

Acc

ura

cy

ideal

INFERENCE PREDICTION

= Process model

= Subject