ACT-R with a physiological substrate

advertisement
1
ACT-R Φ
ACT-R WITH A PHYSIOLOGICAL SUBSTRATE
Christopher L. Dancy
Ph.D. Candidate
Applied Cognitive Science Lab
The College of Information Sciences and Technology
The Pennsylvania State University
“Science involves confronting our `absolute stupidity'. That kind of stupidity is an
existential fact, inherent in our efforts to push our way into the unknown.” – Schwartz,
2008
ACT-R Workshop July 11, 2013
Ways Physiology can Modulate
Cognition and Behavior (some priming…)
2

Appetitive Motivations1,2
 Hunger,



Thirst, Thermal Balance, etc.
Sleep
Stress3
Need to void4
And these all interact!5
1.
2.
3.
4.
5.
Panksepp (2012)
Mogg et al. (1998)
Joëls and Baram (2009)
Tuk et al. (2011)
Montano et al. (2012)
Biology/Physiology in ACT-R
3

Ritter (2007, 2009) – Simulating the effects of stress
& caffeine
Biology/Physiology in ACT-R
4


Ritter (2007, 2009) – Simulating the effects of stress
& caffeine
Changed parameters to simulate participants in
different groups (challenged, threatened, caffeine)
 seconds-per-syllable
(SYL)
 base level constant (BLC)
 activation noise (ANS)

Parameter values were found using GA & were
static across the task
Biology/Physiology in ACT-R
5

Gunzelmann (2009, 2012) - Simulating the effects
of Fatigue/Sleep deprevation
Biology/Physiology in ACT-R
6

Gunzelmann (2009, 2012) - Simulating the effects
of Fatigue/Sleep deprevation
 Used
a model of fatigue (cognitive throughput or
alertness) due to sleep deprevation (CNPA)
 Connected model to DM (activation) and Procedural
(utility)
Questions from the two examples
7



How can we make cognition change physiology too
(and in real-time)?
How can we combine results?
How can we generalize the results?
HumMod1
8
1. Hester et al., 2011
Why HumMod?
9

Integrative model1
 Want
to avoid “micro” computational models of
physiology

Top-down organization
 Provides
macroscopic representation of physiology and
some underlying functionality

Open-source model (XML)
 Allows
verification, validation, and modification (if
needed)

It’s software that works(!!!)
1. Hester et al., 2011
About HumMod
10

Inputs (parameters)
 Exogenous
 e.g.
Epinephrine pump, IV drip
 Modify
autonomic nerve activity
 “Lifestyle”
 e.g.
changes to variables
settings
- Air supply, Exercise, Diet
About HumMod
11
ACT-R
12
1. Anderson et al., 2008
1
ACT-RΦ : An Extension to ACT-R
13
A Modification of a Subtraction Model1,2
14
1. Ritter et al., 2009
2. Dancy et al., Accepted
Subtraction model results1 (n=200)
15
1. Dancy et al., Accepted
Subtraction model results (n = 1,582,000 OR 2 * 3955
* 200)
16
1. Dancy et al., Accepted
Subtraction model results (n=200 * 3955 * 2)
17
1. Dancy et al., Accepted
A Thirsty model
18
1. Dancy et al., 2013
Thirsty model Results
19
Re je ct production rule utility
S u b je c tiv e Th irs t
10
8
6
4
2
0
-1.5
- 1 .0
-0 .5
0 .0
A c c e p t U tility
Decision
Acceptexperiment
Acceptmodel
Rejectexperiment
1. Dancy et al., 2013
Osmolarity (sd)
Not Reported
306.37(0.2)
Not Reported
Subj. Thirst (sd)
8.90(1.7)
7.94(1.2)
5.60(1.6)
Rejectmodel
305.86(3.74)
4.82(1.2)
Bothexperiment
310(5.0)
7.30(1.6)
Bothmodel
306.27(1.7)
7.29(1.7)
0.5
Discussion
Some things for which it could be useful
20

Perseveration & Autonomy

Do you know what Perseveration is? Do you know what Perseveration is?
know what Perseveration is? Do you know what Perseveration is? Do you know what Perseveration is?

Military Simulations
 Energy,
Fatigue, and Stress (Physical and Mental)
Do you
Discussion
Some Possible Road Blocks
21

Visceral Sensory/Perception Conflict
 E.g.



Hunger and Pain
What direct connections should we make?
SPEED!!!
Ease of Use and Visualization
 E.g.,
for 6,000+ variables
Conclusion
A Future for ACT-R (and other architectures)?
22


This approach has several areas of improvement
(see last slide) but that’s expected…
How do we use existing literature to develop these
connections?
 Reviews
of moderators that acknowledge multiple levels
(e.g., Joëls, 2009; Pankepp, 2012) are helpful.
 Experiments that record (& control for) physiological
data with psychological data (e.g., Wright, 2012) are
helpful.
Acknowledgements
23




ACS Lab @ Penn State
Robert Hester
Keith Berry
Jon Morgan
References
24
Anderson, J. R., Fincham, J. M., Qin, Y., & Stocco, A. (2008). A
central circuit of the mind. Trends in Cognitive Sciences, 12(4), 136143.
Dancy, C. L., & Kaulakis, R. (2013). Towards Adding Bottom-Up
Homeostatic Affect to ACT-R. In Proceedings of the Twelfth
International Conference on Cognitive Modeling:Ottawa, Canada.
Dancy, C. L., Ritter, F. E., & Berry, K. (Accepted). Using a
cognitive architecture with a physiological substrate to represent
effects of psychological stress on cognition. Computational &
Mathematical Organization Theory.
Gunzelmann, G., Gluck, K. A., Richard Moore Jr, L., & Dinges, D.
F. (2012). Diminished access to declarative knowledge with sleep
deprivation. Cognitive Systems Research, 13(1), 1-11.
Gunzelmann, G., Gross, J. B., Gluck, K. A., & Dinges, D. F.
(2009). Sleep deprivation and sustained attention performance:
Integrating mathematical and cognitive modeling. Cognitive Science,
33(5), 880-910.
Hester, R. L., Brown, A. J., Husband, L., Iliescu, R., Pruett, D.,
Summers, R., et al. (2011). HumMod: A modeling environment for
the simulation of integrative human physiology. Frontiers in
Physiology, 2(12).
Joëls, M., & Baram, T. Z. (2009). The neuro-symphony of stress.
[10.1038/nrn2632]. Nature Review in Neuroscience, 10(6), 459-466.
Mogg, K., Bradley, B. P., Hyare, H., & Lee, S. (1998). Selective
attention to food-related stimuli in hunger: Are attentional biases
specific to emotional and psychopathological states, or are they also
found in normal drive states? Behaviour Research and Therapy,
36(2), 227-237.
Montano, N., Tobaldini, E., & Porta, A. (2012). The Autonomic
Nervous System Stress Challenges and Immunity in Space. In A.
Chouker (Ed.), (pp. 71-86): Springer Berlin Heidelberg.
Panksepp, J., & Biven, L. (2012). The Archeology of Mind:
Neuroevoloutionary Origins of Human Emotions. New York, NY:
W.W. Norton & Company.
Ritter, F. E., Kase, S. E., Klein, L. C., Bennett, J., & Schoelles, M.
(2009). Fitting a model to behavior tells us what changes cognitively
when under stress and with caffeine. In Proceedings of the the
Biologically Inspired Cognitive Architectures Symposium at the AAAI
Fall Symposium. Keynote presentation, 109-115:Washington, DC.
Ritter, F. E., Reifers, A. L., Klein, L. C., & Schoelles, M. J. (2007).
Lessons from defining theories of stress. In W. D. Gray (Ed.),
Integrated Models of Cognitive Systems (pp. 254-262). New York,
NY: OUP.
Tuk, M. A., Trampe, D., & Warlop, L. (2011). Inhibitory Spillover.
Psychological Science, 22(5), 627-633.
Wright, N. D., Hodgson, K., Fleming, S. M., Symmonds, M.,
Guitart-Masip, M., & Dolan, R. J. (2012). Human responses to
unfairness with primary rewards and their biological limits.
[10.1038/srep00593]. Scientific Reports, 2.
Download