0.25
0.5
0.75
1.25
1.5
1.75
2
Supermodeling: Consensus by Synchronization of Alternative Models
Published on 2011-11-084117 Views
Computational models of an ongoing objective process, as in weather forecasting, must continually assimilate new observational data as they run. Both "truth" and "model" are chaotic systems that thu
Presentation
Consensus By Synchronization of Alternative Models00:00
Data Assimilation06:59:55
Suppose the World is a Lorenz System and Only x is Observed27:27:04
Let a Collection of Models Assimilate Data From (Synchronize With) One Another; Adapt the Coupling Coefficients93:50:39
Summary137:43:45
Coupled Model Intercomparison Project192:47:18
Error in annual mean surface air temperatures217:05:16
Example: Divergent Model Projections of Regional Precipitation Change233:34:02
Test Case: Fusing 3 Lorenz Systems With Different Parameters239:16:36
Supermodeling Relies on 3-Way Synchronization of Truth and Alternative Models329:48:54
...Or Can Use Standard Machine Learning Methods to Adapt Inter-Model Connections (!)350:04:58
...Or Can Use Standard Machine Learning Methods to Adapt Inter-Model Connections (2)361:46:55
Learning Algorithm379:34:58
Supermodeling Works With Multi-time-scale Models401:17:34
What if all models err in the same way?456:39:35
What if all models are biased in same direction?472:08:46
What if parameters shift between training and testing?530:48:48
What if the connection scheme obtained by cost-minimization is only locally optimal?569:12:54
Extension to PDE's: What is the required spatial density of inter-model coupling?598:13:11
What variables should be coupled?662:18:24
Proposed Adaptive Fusion of Two QG Channel Models697:39:17
Models Synchronize With Each Other and With “Truth”736:45:04
... As the Adaptation Procedure Estimates the Intermodel Connection Coefficient c →1/2774:20:23
Limits of Supermodeling778:40:58