Welcome to our TODO list

Toward a first working implementation

In order from most important to least important.

INTEGRATE data.md CONTENTS INTO THIS TODO LIST
Fix line 354 @@@@ nn.go
Finish replacing all mentions of layers, actions, etc with EmergentStack functions
Run the model, ensuring it produces realistic results
Remove all direct references of enviro and intero from Model.go; these should be called via some function that iterates over an array in which they're stored.
Have system for mapping actions/complex actions in NN to deliberate actions in Unity (complex actions will likely raise some issues with our current system)
Implement method for easily switching on and off Requirements condition for updates
Run the model with Unity, ensuring it produces realistic results

Have no "enviro" state updates (like in current Model implementation/diagram); define enviro state updates in Unity. For instance, upon eating food:
- Intero state is updated in neural net model
- Enviro-global state is updated in Unity => Enviro-local state updated in the neural net model
- BUT have it so non-Unity implementation updates Enviro manually, whereas Unity implementation ignores Enviro manual updates
Model Unity and NN message enviro equation
Plan a schematic, git page, etc for the project
Refactored Model.go to use go map when defining parameters, rather than two arrays
Pluralized b and dx values
Mathematically modeled and implemented TimeIncrements
Remove cost function from model. Perhaps we just need the state updates to intrinsically act as the cost (e.g., going on a walk increases need for food and sleep). The neural network model hidden layers will (ideally) take care of representing "cost".
Added units to all variable names
Made layer setup explicit
Removed "current" tensors
Defined EnviroTsr and InteroTsr procedurally based on Model.go code
Created MathFunctions script, mathematiclaly modeled and implemented EnviroUnityToNN
Migrate to leabra v1.2.3, build initial model!