Misc polish

docs/Project.toml

@@ -5,7 +5,6 @@ DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
 Glob = "c27321d9-0574-5035-807b-f59d2c89b15c"
-HierarchicalGaussianFiltering = "63d42c3e-681c-42be-892f-a47f35336a79"
 JuliaFormatter = "98e50ef6-434e-11e9-1051-2b60c6c9e899"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"

docs/src/Using_the_package/Simulation_with_an_agent.jl

@@ -102,7 +102,7 @@ function custom_rescorla_wagner_softmax(agent, input)
 
     ## Read in parameters from the agent
     learning_rate = agent.parameters["learning_rate"]
-    action_precision = agent.parameters["softmax_action_precision"]
+    action_precision = agent.parameters["action_precision"]
 
     ## Read in states with an initial value
     old_value = agent.states["value_binary"]
@@ -174,7 +174,7 @@ end
 
 parameters = Dict(
     "learning_rate" => 1,
-    "softmax_action_precision" => 1,
+    "action_precision" => 1,
     ("initial", "value_cont") => 0,
     ("initial", "value_binary") => 0,
 )

docs/src/Using_the_package/creating_own_action_model.jl

@@ -111,7 +111,7 @@ function custom_rescorla_wagner_softmax(agent, input)
 
     ## Read in parameters from the agent
     learning_rate = agent.parameters["learning_rate"]
-    action_precision = agent.parameters["softmax_action_precision"]
+    action_precision = agent.parameters["action_precision"]
 
     ## Read in states with an initial value
     old_value = agent.states["value"]
@@ -156,7 +156,7 @@ end
 #Set the parameters:
 
 parameters =
-    Dict("learning_rate" => 1, "softmax_action_precision" => 1, ("initial", "value") => 0)
+    Dict("learning_rate" => 1, "action_precision" => 1, ("initial", "value") => 0)
 
 # We set the initial state parameter for "value" state because we need a starting value in the update step. 
 

docs/src/Using_the_package/fitting_an_agent_model_to_data.jl

@@ -104,7 +104,7 @@ get_posteriors(fitted_model)
 
 # Add an extra prior in the Dict
 multiple_priors =
-    Dict("learning_rate" => Normal(1, 0.5), "softmax_action_precision" => Normal(0.8, 0.2))
+    Dict("learning_rate" => Normal(1, 0.5), "action_precision" => Normal(0.8, 0.2))
 
 multiple_fit = fit_model(agent, multiple_priors, inputs, actions)
 

docs/src/Using_the_package/premade_agents_and_models.jl

@@ -62,7 +62,7 @@ set_parameters!(
     agent,
     Dict(
         "learning_rate" => 0.79,
-        "softmax_action_precision" => 0.60,
+        "action_precision" => 0.60,
         ("initial", "value") => 1,
     ),
 )
@@ -74,7 +74,7 @@ agent_custom_parameters = premade_agent(
     "premade_binary_rescorla_wagner_softmax",
     Dict(
         "learning_rate" => 0.7,
-        "softmax_action_precision" => 0.8,
+        "action_precision" => 0.8,
         ("initial", "value") => 1,
     ),
 )

docs/src/Using_the_package/variations_of_util.jl

@@ -48,7 +48,7 @@ get_parameters(agent, [("initial", "value"), "learning_rate"])
 set_parameters!(agent, ("initial", "value"), 1)
 
 # Setting multiple parameters in an agent
-set_parameters!(agent, Dict("learning_rate" => 3, "softmax_action_precision" => 0.5))
+set_parameters!(agent, Dict("learning_rate" => 3, "action_precision" => 0.5))
 
 # See the parameters we have set uising get_parameters function
 get_parameters(agent)
@@ -89,7 +89,7 @@ actions = give_inputs!(agent, inputs)
 # Set a prior for the parameter we wish to fit
 using Distributions
 priors =
-    Dict("softmax_action_precision" => Normal(1, 0.5), "learning_rate" => Normal(1, 0.1))
+    Dict("action_precision" => Normal(1, 0.5), "learning_rate" => Normal(1, 0.1))
 
 # Fit the model
 fitted_model = fit_model(agent, priors, inputs, actions)

docs/src/julia_src_files/creating_own_action_model.jl

@@ -45,32 +45,43 @@ using Plots, StatsPlots
 
 # Rescorla Wagner continuous
 
-function continuous_rescorla_wagner_gaussian(agent, input)
-    #-- Read in parameters and states --#
+
+function continuous_rescorla_wagner_gaussian(agent::Agent, input::Real)
+
+    ## Read in parameters from the agent
     learning_rate = agent.parameters["learning_rate"]
-    action_noise = agent.parameters["action_noise"]
+    action_noise  = agent.parameters["action_noise"]
+
+    ## Read in states with an initial value
     old_value = agent.states["value"]
 
-    #-- Update the Rescolar Wagner's expected value --# 
+    ##We dont have any settings in this model. If we had, we would read them in as well.
+    ##-----This is where the update step starts -------
+
+    ##Get new value state
     new_value = old_value + learning_rate * (input - old_value)
 
-    #-- Sample an action with Gaussian noise --#
-    action_distribution = Distributions.Normal(new_value, action_noise)
 
-    #-- Update states for next round --# 
-    update_states!(agent, "value", new_value)
-    update_states!(agent, "input", new_value)
+    ##-----This is where the update step ends -------
+    ##Create Bernoulli normal distribution our action probability which we calculated in the update step
+    action_distribution = Distributions.Normal(new_value, action_noise)
 
+    ##Update the states and save them to agent's history
+    update_states!(agent, Dict(
+        "value" => new_value,
+        "input" => input,
+    ))
 
-    #-- return the distribution to sample actions from
+    ## return the action distribution to sample actions from
     return action_distribution
 end
 
+
 #-- define parameters and states --#
 parameters = Dict(
     "learning_rate" => 0.8,
     "action_noise" => 1,
-    ("initial", "value") => 0)
+    InitialStateParameter("value") => 0)
 
 states = Dict(
     "value" => missing,
@@ -102,60 +113,36 @@ plot!(actions, linetype = :scatter, label = "action")
 
 
 
-
-
-
-
-plot_trajectory(agent, "action")
-plot_trajectory!(agent, "input")
-
-reset!(agent)
-
-# With binary inputs
-inputs = [0, 1, 0, 0, 1, 1, 1, 0, 0, 1]
-give_inputs!(agent, inputs)
-#-
-plot_trajectory(agent, "action")
-plot(inputs)
-
 # ## A Binary Rescorla-Wagner
+function binary_rescorla_wagner_softmax(agent::Agent, input::Union{Bool,Integer})
 
-function custom_rescorla_wagner_softmax(agent, input)
-
-    ## Read in parameters from the agent
+    #Read in parameters
     learning_rate = agent.parameters["learning_rate"]
-    action_precision = agent.parameters["softmax_action_precision"]
+    action_precision = agent.parameters["action_precision"]
 
-    ## Read in states with an initial value
+    #Read in states
     old_value = agent.states["value"]
 
-    ##We dont have any settings in this model. If we had, we would read them in as well. 
-    ##-----This is where the update step starts ------- 
-
-    ## Sigmoid transform the value
+    #Sigmoid transform the value
     old_value_probability = 1 / (1 + exp(-old_value))
 
-    ##Get new value state
+    #Get new value state
     new_value = old_value + learning_rate * (input - old_value_probability)
 
-    ##Pass through softmax to get action probability
+    #Pass through softmax to get action probability
     action_probability = 1 / (1 + exp(-action_precision * new_value))
 
-    ##-----This is where the update step ends ------- 
-    ##Create Bernoulli normal distribution our action probability which we calculated in the update step
+    #Create Bernoulli normal distribution with mean of the target value and a standard deviation from parameters
     action_distribution = Distributions.Bernoulli(action_probability)
 
-    ##Update the states and save them to agent's history 
+    #Update states
+    update_states!(agent, Dict(
+        "value" => new_value,
+        "value_probability" => 1 / (1 + exp(-new_value)),
+        "action_probability" => action_probability,
+        "input" => input,
+    ))
 
-    agent.states["value"] = new_value
-    agent.states["transformed_value"] = 1 / (1 + exp(-new_value))
-    agent.states["action_probability"] = action_probability
-
-    push!(agent.history["value"], new_value)
-    push!(agent.history["transformed_value"], 1 / (1 + exp(-new_value)))
-    push!(agent.history["action_probability"], action_probability)
-
-    ## return the action distribution to sample actions from
     return action_distribution
 end
 
@@ -169,19 +156,22 @@ end
 #Set the parameters:
 
 parameters =
-    Dict("learning_rate" => 1, "softmax_action_precision" => 1, ("initial", "value") => 0)
+    Dict("learning_rate" => 1,
+     "action_precision" => 1, 
+     InitialStateParameter("value") => 0)
 
 # We set the initial state parameter for "value" state because we need a starting value in the update step. 
 
 # Let us define the states in the agent:
 states = Dict(
     "value" => missing,
-    "transformed_value" => missing,
+    "value_probability" => missing,
     "action_probability" => missing,
+    "input" => missing
 )
 
 # And lastly the action model:
-action_model = custom_rescorla_wagner_softmax
+action_model = binary_rescorla_wagner_softmax
 
 # We can now initialize our agent with the action model, parameters and states.
 agent = init_agent(action_model, parameters = parameters, states = states)

docs/src/julia_src_files/fitting_an_agent_model_to_data.jl

@@ -104,7 +104,7 @@ get_posteriors(fitted_model)
 
 # Add an extra prior in the Dict
 multiple_priors =
-    Dict("learning_rate" => Normal(1, 0.5), "softmax_action_precision" => Normal(0.8, 0.2))
+    Dict("learning_rate" => Normal(1, 0.5), "action_precision" => Normal(0.8, 0.2))
 
 multiple_fit = fit_model(agent, multiple_priors, inputs, actions)
 

docs/src/julia_src_files/premade_agents_and_models.jl

@@ -62,7 +62,7 @@ set_parameters!(
     agent,
     Dict(
         "learning_rate" => 0.79,
-        "softmax_action_precision" => 0.60,
+        "action_precision" => 0.60,
         ("initial", "value") => 1,
     ),
 )
@@ -74,7 +74,7 @@ agent_custom_parameters = premade_agent(
     "binary_rescorla_wagner_softmax",
     Dict(
         "learning_rate" => 0.7,
-        "softmax_action_precision" => 0.8,
+        "action_precision" => 0.8,
         ("initial", "value") => 1,
     ),
 )

docs/src/julia_src_files/simulation_with_an_agent.jl

@@ -102,7 +102,7 @@ function custom_rescorla_wagner_softmax(agent, input)
 
     ## Read in parameters from the agent
     learning_rate = agent.parameters["learning_rate"]
-    action_precision = agent.parameters["softmax_action_precision"]
+    action_precision = agent.parameters["action_precision"]
 
     ## Read in states with an initial value
     old_value = agent.states["value_binary"]
@@ -174,7 +174,7 @@ end
 
 parameters = Dict(
     "learning_rate" => 1,
-    "softmax_action_precision" => 1,
+    "action_precision" => 1,
     ("initial", "value_cont") => 0,
     ("initial", "value_binary") => 0,
 )

docs/src/julia_src_files/variations_of_util.jl

@@ -48,7 +48,7 @@ get_parameters(agent, [("initial", "value"), "learning_rate"])
 set_parameters!(agent, ("initial", "value"), 1)
 
 # Setting multiple parameters in an agent
-set_parameters!(agent, Dict("learning_rate" => 3, "softmax_action_precision" => 0.5))
+set_parameters!(agent, Dict("learning_rate" => 3, "action_precision" => 0.5))
 
 # See the parameters we have set uising get_parameters function
 get_parameters(agent)
@@ -89,7 +89,7 @@ actions = give_inputs!(agent, inputs)
 # Set a prior for the parameter we wish to fit
 using Distributions
 priors =
-    Dict("softmax_action_precision" => Normal(1, 0.5), "learning_rate" => Normal(1, 0.1))
+    Dict("action_precision" => Normal(1, 0.5), "learning_rate" => Normal(1, 0.1))
 
 # Fit the model
 fitted_model = fit_model(agent, priors, inputs, actions)

docs/src/markdowns/creating_own_action_model.md

@@ -45,7 +45,7 @@ function custom_rescorla_wagner_softmax(agent::Agent, input)
 
     # Read in parameters from the agent
     learning_rate = agent.parameters["learning_rate"]
-    action_precision = agent.parameters["softmax_action_precision"]
+    action_precision = agent.parameters["action_precision"]
 
     # Read in states with an initial value
     old_value = agent.states["value"]
@@ -90,7 +90,7 @@ We can define the agent now. Let us do it with the init_agent() function. We nee
 
 parameters = Dict(
     "learning_rate" => 1,
-    "softmax_action_precision" => 1,
+    "action_precision" => 1,
     ("initial", "value") => 0,)
 ````
 

docs/src/markdowns/fitting_an_agent_model_to_data.md

@@ -118,7 +118,7 @@ By adding multiple parameter priors you can autimatically fit them with fit\_mod
 Add an extra prior in the Dict
 
 ````@example fitting_an_agent_model_to_data
-multiple_priors = Dict("learning_rate" => Normal(1, 0.5),"softmax_action_precision"=> Normal(0.8,0.2))
+multiple_priors = Dict("learning_rate" => Normal(1, 0.5),"action_precision"=> Normal(0.8,0.2))
 
 multiple_fit = fit_model(agent, multiple_priors, inputs, actions)
 ````

docs/src/markdowns/premade_agents_and_models.md

@@ -79,7 +79,7 @@ If you wish to change multiple parameters in the agent, you can define a dict of
 
 ````@example premade_agents_and_models
 set_parameters!(agent, Dict("learning_rate" => 0.79,
-                        "softmax_action_precision" => 0.60,
+                        "action_precision" => 0.60,
                         ("initial", "value") => 1))
 ````
 
@@ -89,7 +89,7 @@ If you know which parameter values you wish to use when defining your agent, you
 
 ````@example premade_agents_and_models
 agent_custom_parameters = premade_agent("premade_binary_rescorla_wagner_softmax", Dict("learning_rate" => 0.7,
-                                        "softmax_action_precision" => 0.8,
+                                        "action_precision" => 0.8,
                                         ("initial", "value") => 1)
 )
 

docs/src/markdowns/variations_of_util.md

@@ -75,7 +75,7 @@ set_parameters!(agent,("initial", "value"), 1 )
 Setting multiple parameters in an agent
 
 ````@example variations_of_util
-set_parameters!(agent, Dict("learning_rate" => 3, "softmax_action_precision"=>0.5))
+set_parameters!(agent, Dict("learning_rate" => 3, "action_precision"=>0.5))
 ````
 
 See the parameters we have set uising get_parameters function
@@ -136,7 +136,7 @@ Set a prior for the parameter we wish to fit
 
 ````@example variations_of_util
 using Distributions
-priors = Dict("softmax_action_precision" => Normal(1, 0.5), "learning_rate"=> Normal(1, 0.1))
+priors = Dict("action_precision" => Normal(1, 0.5), "learning_rate"=> Normal(1, 0.1))
 ````
 
 Fit the model

src/ActionModels.jl

@@ -4,13 +4,14 @@ module ActionModels
 using ReverseDiff, Turing, Distributions, DataFrames, RecipesBase, Logging, Distributed
 using Turing: DynamicPPL, AutoReverseDiff
 #Export functions
-export Agent, RejectParameters, SharedParameter, Multilevel
+export Agent, RejectParameters, GroupedParameters, Multilevel
 export init_agent, premade_agent, warn_premade_defaults, multiple_actions, check_agent
 export create_agent_model, fit_model
 export plot_parameter_distribution,
     plot_predictive_simulation, plot_trajectory, plot_trajectory!
 export get_history, get_states, get_parameters, set_parameters!, reset!, give_inputs!, single_input!
 export get_posteriors, update_states!, set_save_history!
+export InitialStateParameter, ParameterGroup
 
 #Load premade agents
 function __init__()