Thermodynamics as a theory of bounded rational decision-making thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Thermodynamics as a theory of bounded rational decision-making

Published on Oct 16, 20124660 Views

Perfectly rational decision-makers maximize expected utility, but crucially ignore the resource costs incurred when determining optimal actions. Here we propose an information-theoretic formalization

Related categories

Chapter list

Free Energy & Bounded Rationality00:00
Introduction - 100:24
Introduction - 200:42
Caveat: Metareasoning does not work! - 100:59
Caveat: Metareasoning does not work! - 201:47
Caveat: Metareasoning does not work! - 302:04
Caveat: Metareasoning does not work! - 402:08
Bounded Rationality - 102:41
Bounded Rationality - 202:51
Bounded Rationality - 303:30
The Cost of Transformations - 103:44
The Cost of Transformations - 204:11
The Cost of Transformations - 305:12
The Model of Information State05:24
Measure-Theoretic Formalization of Transformations08:16
Axioms of Transformation Costs - 109:24
Axioms of Transformation Costs - 209:49
Measure-Theoretic Formalization of Decisions10:21
Decisions11:28
Decisions (cont.)11:39
Equilibrium Distribution11:52
NFED Extremum12:50
Operational Interpretation of Inverse Temperature14:16
Decision Trees15:37
Goal: Generalized Decision Trees17:04
Change of Temperature17:34
Construction of Generalized Decision Trees18:00
Generalized Optimality Equations18:32
Conclusions18:40
The free energy principle in human sensorimotor control18:44
Risk in Decision-Making18:46
Motor Control and Maximum Expected Gain - 118:53
Motor Control and Maximum Expected Gain - 219:14
Optimal Feedback Control19:19
Variational Principle19:34
Certainty-equivalent19:37
Equilibrium distribution - 119:41
Equilibrium distribution - 219:50
Risk-sensitivity and model uncertainty20:25
Experimental Studies21:17
Study 121:28
Study 1: Experimental Setup21:29
Study 1: Model fit22:52
Study 1: Results23:11
Study 323:55
Study 3: Experimental Setup23:57
Study 3: Bayesian Sensorimotor Integration24:58
Study 3: Model Prediction26:10
Study 3: Results - 126:11
Study 3: Results - 228:02
Conclusion28:03