Files
Ironpen/src/types.jl

771 lines
34 KiB
Julia
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
module types
export
# struct
IronpenStruct, model, knowledgeFn, lif_neuron, alif_neuron, linear_neuron,
kfn_1, compute_neuron, neuron, output_neuron, passthrough_neuron,
# function
instantiate_custom_types, init_neuron, populate_neuron,
add_neuron!
using Random, Flux, LinearAlgebra
#------------------------------------------------------------------------------------------------100
abstract type Ironpen end
abstract type knowledgeFn <: Ironpen end
abstract type neuron <: Ironpen end
abstract type input_neuron <: neuron end
abstract type output_neuron <: neuron end
abstract type compute_neuron <: neuron end
#------------------------------------------------------------------------------------------------100
""" Model struct
"""
Base.@kwdef mutable struct model <: Ironpen
knowledgeFn::Union{Dict,Nothing} = nothing
modelParams::Union{Dict,Nothing} = nothing
error::Union{Float64,Nothing} = 0.0
outputError::Union{Array,Nothing} = Vector{AbstractFloat}()
""" "inference" = no learning params will be collected.
"learning" = neuron will accumulate epsilon_j, compute Δw_rec_change each time
correct answer is available then merge Δw_rec_change into wRecChange then
reset epsilon_j.
"reflect" = neuron will merge wRecChange into w_rec then reset wRecChange. """
learningStage::String = "inference"
softreset::Bool = false
timeStep::Number = 0.0
end
""" Model outer constructor
# Example
I_kfnparams = Dict(
:type => "lif_neuron",
:v_t1 => 0.0, # neuron membrane potential at time = t+1
:v_th => 2.0, # neuron firing threshold (this value is treated as maximum bound if I use auto generate)
:z_t => false, # neuron firing status at time = t
:z_t1 => false, # neuron firing status at time = t+1
:gammaPd => 0.3, # discount factor. The value is from the paper
:phi => 0.0, # psuedo derivative
:refractoryDuration => 2.0, # neuron refractory period in tick
:delta => 1.0,
:tau_m => 20.0, # membrane time constant in millisecond. The value is from the paper
:eta => 0.01, # learning rate
I_kfn = Ironpen_ai_gpu.knowledgeFn(I_kfnparams, lif_neuron_params, alif_neuron_params,
linear_neuron_params)
modelParams_1 = Dict(:knowledgeFn => Dict(:I => I_kfn,
:run => run_kfn),
:learningStage => "doing_inference",)
model_1 = Ironpen_ai_gpu.model(modelParams_1)
"""
function model(params::Dict)
m = model()
m.modelParams = params
fields = fieldnames(typeof(m))
for i in fields
if i in keys(params)
m.:($i) = params[i] # assign params to n struct fields
end
end
return m
end
#------------------------------------------------------------------------------------------------100
""" knowledgeFn struct
"""
Base.@kwdef mutable struct kfn_1 <: knowledgeFn
knowledgeFnName::Union{String,Nothing} = nothing
kfnParams::Union{Dict,Nothing} = nothing # store params of knowledgeFn itself for later use
timeStep::Number = 0.0
# Bn contain error coefficient for both neurons and output neurons in one place
Bn::Vector{Float64} = Vector{Float64}() # error projection coefficient from kfn output's error to each neurons's error
neuronsArray::Union{Array,Nothing} = [] # put neurons here
""" put output neuron here. I seperate output neuron because
1. its calculation is difference than other neuron types
2. other neuron type will not induced to connnect to output neuron
3. output neuron does not induced to connect to its own type """
outputNeuronsArray::Union{Array,Nothing} = []
""" "inference" = no learning params will be collected.
"learning" = neuron will accumulate epsilon_j, compute Δw_rec_change each time
correct answer is available then merge Δw_rec_change into wRecChange then
reset epsilon_j.
"reflect" = neuron will merge wRecChange into w_rec then reset wRecChange. """
learningStage::String = "inference"
error::Union{Float64,Nothing} = nothing
softreset::Bool = false
firedNeurons::Array{Int64} = Vector{Int64}() # store unique id of firing neurons to be used when random neuron connection
firedNeurons_t0::Union{Vector{Bool},Nothing} = nothing # store firing state of all neurons at t0
firedNeurons_t1::Union{Vector{Bool},Nothing} = nothing # store firing state of all neurons at t1
avgNeuronsFiringRate::Union{Float64,Nothing} = 0.0 # for displaying average firing rate over all neurons
avgNeurons_v_t1::Union{Float64,Nothing} = 0.0 # for displaying average v_t1 over all neurons
end
#------------------------------------------------------------------------------------------------100
""" Knowledge function outer constructor >>> auto generate <<<
# Example
lif_neuron_params = Dict(
:type => "lif_neuron",
:v_th => 1.2, # neuron firing threshold (this value is treated as maximum bound if I use auto generate)
:z_t => false, # neuron firing status at time = t
:gammaPd => 0.3, # discount factor. The value is from the paper
:refractoryDuration => 2.0, # neuron refractory period in tick
:delta => 1.0,
:tau_m => 5.0, # membrane time constant in millisecond. It should equals to time use for 1 sequence
)
alif_neuron_params = Dict(
:type => "alif_neuron",
:v_th => 1.2, # neuron firing threshold (this value is treated as maximum bound if I use auto generate)
:z_t => false, # neuron firing status at time = t
:gammaPd => 0.3, # discount factor. The value is from the paper
:refractoryDuration => 2.0, # neuron refractory period in millisecond
:delta => 1.0,
:tau_m => 5.0, # membrane time constant in millisecond. It should equals to time use for 1 sequence
# adaptation time constant in millisecond. It should equals to total time SNN takes to
# perform a task i.e. equals to episode length
:tau_a => 10.0,
:beta => 0.15, # constant.
:a => 0.0,
)
linear_neuron_params = Dict(
:type => "linear_neuron",
:k => 0.9, # output leakink coefficient
:tau_out => 5.0, # output time constant in millisecond. It should equals to time use for 1 sequence
:out => 0.0, # neuron's output value store here
)
I_kfnparams = Dict(
:knowledgeFnName => "I",
:lif_neuron_number => 200,
:alif_neuron_number => 100, # from Allen Institute, ALIF is 40% of LIF
:linear_neuron_number => 5, # output neuron, this is also the output length
:Bn => "random", # error projection coefficient from kfn output's error to each neurons's error
:learning_rate => 0.01,
:neuron_connection_pattern => "100%", # number of each neuron subscribe to other neuron in knowledgeFn.neuronsArray
:output_neuron_connection_pattern => "100%", # "60%" of kfn.neuronsArray or number
:maximum_input_data_length => 5, # in case of GloVe word encoding, it is 300
:neuron_w_in_generation_pattern => "random", # number or "random"
:neuron_w_rec_generation_pattern => "random",
:neuron_v_t_default => 0.5,
:neuron_voltage_drop_percentage => "100%",
:neuron_firing_rate_target => 50.0,
:neuron_learning_rate => 0.01,
:neuron_c_reg => 0.0001,
:neuron_c_reg_v => 0.0001,
:neuron_optimiser => "ADAM",
:meta_params => Dict(:is_first_cycle => true,
:launch_time => 0.0,))
kfn1 = knowledgeFn(kfnParams, lif_neuron_params, alif_neuron_params, linear_neuron_params)
"""
function kfn_1(kfnParams::Dict)
kfn = kfn_1()
kfn.kfnParams = kfnParams
kfn.knowledgeFnName = kfn.kfnParams[:knowledgeFnName]
if kfn.kfnParams[:compute_neuron_number] < kfn.kfnParams[:total_input_port]
throw(error("number of compute neuron must be greater than input neuron"))
end
# Bn
if kfn.kfnParams[:Bn] == "random"
kfn.Bn = [Random.rand(0:0.001:1) for i in 1:kfn.kfnParams[:compute_neuron_number]]
else # in case I want to specify manually
kfn.Bn = [kfn.kfnParams[:Bn] for i in 1:kfn.kfnParams[:compute_neuron_number]]
end
# assign neurons ID by their position in kfn.neurons array because I think it is
# straight forward way
# add input port
for (k, v) in kfn.kfnParams[:input_port]
current_type = kfn.kfnParams[:input_port][k]
for i = 1:current_type[:numbers]
n_id = length(kfn.neuronsArray) + 1
neuron = init_neuron(n_id, current_type[:params], kfn.kfnParams)
push!(kfn.neuronsArray, neuron)
end
end
# add compute neurons
for (k, v) in kfn.kfnParams[:compute_neuron]
current_type = kfn.kfnParams[:compute_neuron][k]
for i = 1:current_type[:numbers]
n_id = length(kfn.neuronsArray) + 1
neuron = init_neuron(n_id, current_type[:params], kfn.kfnParams)
push!(kfn.neuronsArray, neuron)
end
end
for i = 1:kfn.kfnParams[:output_port][:numbers]
neuron = init_neuron(i, kfn.kfnParams[:output_port][:params],
kfn.kfnParams)
push!(kfn.outputNeuronsArray, neuron)
end
for n in kfn.neuronsArray
if typeof(n) <: compute_neuron
n.firingRateTarget = kfn.kfnParams[:neuron_firing_rate_target]
end
end
# excitatory neuron to inhabitory neuron = 60:40 % of compute_neuron
ex_number = Int(floor(0.6 * kfn.kfnParams[:compute_neuron_number]))
ex_n = [1 for i in 1:ex_number]
in_number = kfn.kfnParams[:compute_neuron_number] - ex_number
in_n = [-1 for i in 1:in_number]
ex_in = shuffle!([ex_n; in_n])
# input neurons are always excitatory, compute_neurons are random between excitatory
# and inhabitory
for n in reverse(kfn.neuronsArray)
try n.ExInType = pop!(ex_in) catch end
end
# add ExInType into each compute_neuron subExInType
for n in reverse(kfn.neuronsArray)
try # input neuron doest have n.subscriptionList
for sub_id in n.subscriptionList
n_ExInType = kfn.neuronsArray[sub_id].ExInType
push!(n.subExInType, n_ExInType)
end
catch
end
end
# add ExInType into each output neuron subExInType
for n in kfn.outputNeuronsArray
try # input neuron doest have n.subscriptionList
for sub_id in n.subscriptionList
n_ExInType = kfn.neuronsArray[sub_id].ExInType
push!(n.subExInType, n_ExInType)
end
catch
end
end
return kfn
end
#------------------------------------------------------------------------------------------------100
""" passthrough_neuron struct
"""
Base.@kwdef mutable struct passthrough_neuron <: input_neuron
id::Union{Int64,Nothing} = nothing # ID of this neuron which is it position in knowledgeFn array
type::String = "passthrough_neuron"
knowledgeFnName::Union{String,Nothing} = nothing # knowledgeFn that this neuron belongs to
z_t::Bool = false
z_t1::Bool = false
timeStep::Number = 0.0 # current time
ExInType::Integer = 1 # 1 excitatory, -1 inhabitory. input neuron is always excitatory
end
function passthrough_neuron(params::Dict)
n = passthrough_neuron()
field_names = fieldnames(typeof(n))
for i in field_names
if i in keys(params)
if i == :optimiser
opt_type = string(split(params[i], ".")[end])
n.:($i) = load_optimiser(opt_type)
else
n.:($i) = params[i] # assign params to n struct fields
end
end
end
return n
end
#------------------------------------------------------------------------------------------------100
""" lif_neuron struct
"""
Base.@kwdef mutable struct lif_neuron <: compute_neuron
id::Union{Int64,Nothing} = nothing # this neuron ID i.e. position of this neuron in knowledgeFn
type::String = "lif_neuron"
ExInType::Integer = 1 # 1 excitatory, -1 inhabitory
# Bn::Union{Float64,Nothing} = Random.rand() # Bias for neuron error
knowledgeFnName::Union{String,Nothing} = nothing # knowledgeFn that this neuron belongs to
subscriptionList::Union{Array{Int64},Nothing} = nothing # list of other neuron that this neuron synapse subscribed to
subExInType::Array{Int64} = Vector{Int64}() # store ExIn type of subscribed neurons
timeStep::Number = 0.0 # current time
w_rec::Union{Array{Float64},Nothing} = nothing # synaptic weight (for receiving signal from other neuron)
v_t::Float64 = rand() # vᵗ, postsynaptic neuron membrane potential of previous timestep
v_t1::Float64 = 0.0 # vᵗ⁺¹, postsynaptic neuron membrane potential at current timestep
v_t_default::Union{Float64,Nothing} = 0.0 # default membrane potential voltage
v_th::Float64 = 1.0 # vᵗʰ, neuron firing threshold
vRest::Float64 = 0.0 # resting potential after neuron fired
z_t::Bool = false # zᵗ, neuron postsynaptic firing of previous timestep
# zᵗ⁺¹, neuron firing status at time = t+1. I need this because the way I calculate all
# neurons forward function at each timestep-by-timestep is to do every neuron
# forward calculation. Each neuron requires access to other neuron's firing status
# during v_t1 calculation hence I need a variable to hold z_t1 so that I'm not replacing z_t
z_t1::Bool = false # neuron postsynaptic firing at current timestep (after neuron's calculation)
z_i_t::Union{Array{Bool},Nothing} = nothing # neuron presynaptic firing at current timestep (which is other neuron postsynaptic firing of previous timestep)
synapticStrength::Union{Array{Float64},Nothing} = nothing
synapticStrengthLimit::Union{NamedTuple,Nothing} = (lowerlimit=(0=>0), upperlimit=(10=>10))
gammaPd::Union{Float64,Nothing} = 0.3 # γ_pd, discount factor, value from paper
alpha::Union{Float64,Nothing} = nothing # α, neuron membrane potential decay factor
phi::Union{Float64,Nothing} = nothing # ϕ, psuedo derivative
epsilonRec::Union{Array{Float64},Nothing} = nothing # ϵ_rec, eligibility vector for neuron spike
decayedEpsilonRec::Union{Array{Float64},Nothing} = nothing # α * epsilonRec
eRec::Union{Array{Float64},Nothing} = nothing # eligibility trace for neuron spike
delta::Union{Float64,Nothing} = 1.0 # δ, discreate timestep size in millisecond
refractoryDuration::Union{Float64,Nothing} = 3 # neuron's refratory period in millisecond
# refractory_state_active::Union{Bool,Nothing} = false # if true, neuron is in refractory state and cannot process new information
refractoryCounter::Integer = 0
tau_m::Union{Float64,Nothing} = nothing # τ_m, membrane time constant in millisecond
eta::Union{Float64,Nothing} = 0.01 # η, learning rate
wRecChange::Union{Array{Float64},Nothing} = nothing # Δw_rec, cumulated w_rec change
recSignal::Union{Float64,Nothing} = nothing # incoming recurrent signal
alpha_v_t::Union{Float64,Nothing} = nothing # alpha * v_t
error::Union{Float64,Nothing} = nothing # local neuron error
optimiser::Union{Any,Nothing} = load_optimiser("AdaBelief") # Flux optimizer
firingCounter::Integer = 0 # store how many times neuron fires
firingRateTarget::Float64 = 20.0 # neuron's target firing rate in Hz
firingDiff::Float64 = 0.0 # e-prop supplement paper equation 5
firingRateError::Float64 = 0.0 # local neuron error w.r.t. firing regularization
firingRate::Float64 = 0.0 # running average of firing rate in Hz
""" "inference" = no learning params will be collected.
"learning" = neuron will accumulate epsilon_j, compute Δw_rec_change each time
correct answer is available then merge Δw_rec_change into wRecChange then
reset epsilon_j.
"reflect" = neuron will merge wRecChange into w_rec then reset wRecChange. """
learningStage::String = "inference"
end
""" lif neuron outer constructor
# Example
lif_neuron_params = Dict(
:type => "lif_neuron",
:v_th => 1.2, # neuron firing threshold (this value is treated as maximum bound if I use auto generate)
:z_t => false, # neuron firing status at time = t
:gammaPd => 0.3, # discount factor. The value is from the paper
:refractoryDuration => 2.0, # neuron refractory period in tick
:delta => 1.0,
:tau_m => 5.0, # membrane time constant in millisecond. It should equals to time use for 1 sequence
)
neuron1 = lif_neuron(lif_neuron_params)
"""
function lif_neuron(params::Dict)
n = lif_neuron()
field_names = fieldnames(typeof(n))
for i in field_names
if i in keys(params)
if i == :optimiser
opt_type = string(split(params[i], ".")[end])
n.:($i) = load_optimiser(opt_type)
else
n.:($i) = params[i] # assign params to n struct fields
end
end
end
return n
end
#------------------------------------------------------------------------------------------------100
""" alif_neuron struct
"""
Base.@kwdef mutable struct alif_neuron <: compute_neuron
id::Union{Int64,Nothing} = nothing # this neuron ID i.e. position of this neuron in knowledgeFn
type::String = "alif_neuron"
ExInType::Integer = -1 # 1 excitatory, -1 inhabitory
# Bn::Union{Float64,Nothing} = Random.rand() # Bias for neuron error
knowledgeFnName::Union{String,Nothing} = nothing # knowledgeFn that this neuron belongs to
subscriptionList::Union{Array{Int64},Nothing} = nothing # list of other neuron that this neuron synapse subscribed to
subExInType::Array{Int64} = Vector{Int64}() # store ExIn type of subscribed neurons
timeStep::Union{Number,Nothing} = nothing # current time
w_rec::Union{Array{Float64},Nothing} = nothing # synaptic weight (for receiving signal from other neuron)
v_t::Float64 = rand() # vᵗ, postsynaptic neuron membrane potential of previous timestep
v_t1::Float64 = 0.0 # vᵗ⁺¹, postsynaptic neuron membrane potential at current timestep
v_t_default::Union{Float64,Nothing} = 0.0
v_th::Float64 = 1.0 # vᵗʰ, neuron firing threshold
vRest::Float64 = 0.0 # resting potential after neuron fired
z_t::Bool = false # zᵗ, neuron postsynaptic firing of previous timestep
# zᵗ⁺¹, neuron firing status at time = t+1. I need this because the way I calculate all
# neurons forward function at each timestep-by-timestep is to do every neuron
# forward calculation. Each neuron requires access to other neuron's firing status
# during v_t1 calculation hence I need a variable to hold z_t1 so that I'm not replacing z_t
z_t1::Bool = false # neuron postsynaptic firing at current timestep (after neuron's calculation)
z_i_t::Union{Array{Bool},Nothing} = nothing # neuron presynaptic firing at current timestep (which is other neuron postsynaptic firing of previous timestep)
synapticStrength::Union{Array{Float64},Nothing} = nothing
synapticStrengthLimit::Union{NamedTuple,Nothing} = (lowerlimit=(-5=>0), upperlimit=(5=>5))
alpha::Union{Float64,Nothing} = nothing # α, neuron membrane potential decay factor
delta::Union{Float64,Nothing} = 1.0 # δ, discreate timestep size in millisecond
epsilonRec::Union{Array{Float64},Nothing} = nothing # ϵ_rec(v), eligibility vector for neuron i spike
epsilonRecA::Union{Array{Float64},Nothing} = nothing # ϵ_rec(a)
decayedEpsilonRec::Union{Array{Float64},Nothing} = nothing # α * epsilonRec
eRec_v::Union{Array{Float64},Nothing} = nothing # a component of neuron's eligibility trace resulted from v_t
eRec_a::Union{Array{Float64},Nothing} = nothing # a component of neuron's eligibility trace resulted from av_th
eRec::Union{Array{Float64},Nothing} = nothing # neuron's eligibility trace
eta::Union{Float64,Nothing} = 0.01 # eta, learning rate
gammaPd::Union{Float64,Nothing} = 0.3 # γ_pd, discount factor, value from paper
phi::Union{Float64,Nothing} = nothing # ϕ, psuedo derivative
refractoryDuration::Union{Float64,Nothing} = 3 # neuron's refractory period in millisecond
# refractory_state_active::Union{Bool,Nothing} = false # if true, neuron is in refractory state and cannot process new information
refractoryCounter::Integer = 0
tau_m::Union{Float64,Nothing} = nothing # τ_m, membrane time constant in millisecond
wRecChange::Union{Array{Float64},Nothing} = nothing # Δw_rec, cumulated w_rec change
recSignal::Union{Float64,Nothing} = nothing # incoming recurrent signal
alpha_v_t::Union{Float64,Nothing} = nothing # alpha * v_t
error::Union{Float64,Nothing} = nothing # local neuron error
optimiser::Union{Any,Nothing} = load_optimiser("AdaBelief") # Flux optimizer
firingCounter::Integer = 0 # store how many times neuron fires
firingRateTarget::Float64 = 20.0 # neuron's target firing rate in Hz
firingDiff::Float64 = 0.0 # e-prop supplement paper equation 5
firingRateError::Float64 = 0.0 # local neuron error w.r.t. firing regularization
firingRate::Float64 = 0.0 # running average of firing rate, Hz
tau_a::Union{Float64,Nothing} = nothing # τ_a, adaption time constant in millisecond
beta::Union{Float64,Nothing} = 0.15 # β, constant, value from paper
rho::Union{Float64,Nothing} = nothing # ρ, threshold adaptation decay factor
a::Union{Float64,Nothing} = 0.0 # threshold adaptation
av_th::Union{Float64,Nothing} = nothing # adjusted neuron firing threshold
""" "inference" = no learning params will be collected.
"learning" = neuron will accumulate epsilon_j, compute Δw_rec_change each time
correct answer is available then merge Δw_rec_change into wRecChange then
reset epsilon_j.
"reflect" = neuron will merge wRecChange into w_rec then reset wRecChange. """
learningStage::String = "inference"
end
""" alif neuron outer constructor
# Example
alif_neuron_params = Dict(
:type => "alif_neuron",
:v_th => 1.2, # neuron firing threshold (this value is treated as maximum bound if I
use auto generate)
:z_t => false, # neuron firing status at time = t
:gammaPd => 0.3, # discount factor. The value is from the paper
:refractoryDuration => 2.0, # neuron refractory period in millisecond
:delta => 1.0,
:tau_m => 5.0, # membrane time constant in millisecond. It should equals to time use
for 1 sequence
# adaptation time constant in millisecond. It should equals to total time SNN takes to
# perform a task i.e. equals to episode length
:tau_a => 10.0,
:beta => 0.15, # constant.
:a => 0.0,
)
neuron1 = alif_neuron(alif_neuron_params)
"""
function alif_neuron(params::Dict)
n = alif_neuron()
field_names = fieldnames(typeof(n))
for i in field_names
if i in keys(params)
if i == :optimiser
opt_type = string(split(params[i], ".")[end])
n.:($i) = load_optimiser(opt_type)
else
n.:($i) = params[i] # assign params to n struct fields
end
end
end
return n
end
#------------------------------------------------------------------------------------------------100
""" linear_neuron struct
"""
Base.@kwdef mutable struct linear_neuron <: output_neuron
id::Union{Int64,Nothing} = nothing # ID of this neuron which is it position in knowledgeFn array
type::String = "linear_neuron"
knowledgeFnName::Union{String,Nothing} = nothing # knowledgeFn that this neuron belongs to
subscriptionList::Union{Array{Int64},Nothing} = nothing # list of other neuron that this neuron synapse subscribed to
timeStep::Union{Number,Nothing} = nothing # current time
subExInType::Array{Int64} = Vector{Int64}() # store ExIn type of subscribed neurons
w_rec::Union{Array{Float64},Nothing} = nothing # synaptic weight (for receiving signal from other neuron)
v_t::Float64 = 0.0 # vᵗ, postsynaptic neuron membrane potential of previous timestep
v_t1::Float64 = 0.0 # vᵗ⁺¹, postsynaptic neuron membrane potential at current timestep
v_t_default::Union{Float64,Nothing} = 0.0 # default membrane potential voltage
v_th::Float64 = 1.0 # vᵗʰ, neuron firing threshold
vRest::Float64 = 0.0 # resting potential after neuron fired
# zᵗ⁺¹, neuron firing status at time = t+1. I need this because the way I calculate all
# neurons forward function at each timestep-by-timestep is to do every neuron
# forward calculation. Each neuron requires access to other neuron's firing status
# during v_t1 calculation hence I need a variable to hold z_t1 so that I'm not replacing z_t
z_t1::Bool = false # neuron postsynaptic firing at current timestep (after neuron's calculation)
# neuron presynaptic firing at current timestep (which is other neuron postsynaptic firing of
# previous timestep)
z_i_t::Union{Array{Bool},Nothing} = nothing
synapticStrength::Union{Array{Float64},Nothing} = nothing
synapticStrengthLimit::Union{NamedTuple,Nothing} = (lowerlimit=(-5=>0), upperlimit=(5=>5))
gammaPd::Union{Float64,Nothing} = 0.3 # γ_pd, discount factor, value from paper
alpha::Union{Float64,Nothing} = nothing # α, neuron membrane potential decay factor
phi::Union{Float64,Nothing} = nothing # ϕ, psuedo derivative
epsilonRec::Union{Array{Float64},Nothing} = nothing # ϵ_rec, eligibility vector for neuron spike
decayedEpsilonRec::Union{Array{Float64},Nothing} = nothing # α * epsilonRec
eRec::Union{Array{Float64},Nothing} = nothing # eligibility trace for neuron spike
delta::Union{Float64,Nothing} = 1.0 # δ, discreate timestep size in millisecond
refractoryDuration::Union{Float64,Nothing} = 3 # neuron's refratory period in millisecond
refractoryCounter::Integer = 0
tau_out::Union{Float64,Nothing} = nothing # τ_out, membrane time constant in millisecond
eta::Union{Float64,Nothing} = 0.01 # η, learning rate
wRecChange::Union{Array{Float64},Nothing} = nothing # Δw_rec, cumulated w_rec change
recSignal::Union{Float64,Nothing} = nothing # incoming recurrent signal
alpha_v_t::Union{Float64,Nothing} = nothing # alpha * v_t
error::Union{Float64,Nothing} = nothing # local neuron error
firingCounter::Integer = 0 # store how many times neuron fires
end
""" linear neuron outer constructor
# Example
linear_neuron_params = Dict(
:type => "linear_neuron",
:k => 0.9, # output leakink coefficient
:tau_out => 5.0, # output time constant in millisecond. It should equals to time use for 1 sequence
:out => 0.0, # neuron's output value store here
)
neuron1 = linear_neuron(linear_neuron_params)
"""
function linear_neuron(params::Dict)
n = linear_neuron()
field_names = fieldnames(typeof(n))
for i in field_names
if i in keys(params)
if i == :optimiser
opt_type = string(split(params[i], ".")[end])
n.:($i) = load_optimiser(opt_type)
else
n.:($i) = params[i] # assign params to n struct fields
end
end
end
return n
end
#------------------------------------------------------------------------------------------------100
function load_optimiser(optimiser_name::String; params::Union{Dict,Nothing} = nothing)
if optimiser_name == "AdaBelief"
params = (0.01, (0.9, 0.8))
return Flux.Optimise.AdaBelief(params...)
elseif optimiser_name == "AdaBelief2"
# output neuron requires slower change pace so η is lower than compute neuron at 0.007
# because if w_out change too fast, compute neuron will not able to
# grapse output neuron moving direction i.e. both compute neuron's direction and
# output neuron direction are out of sync.
params = (0.007, (0.9, 0.8))
return Flux.Optimise.AdaBelief(params...)
else
error("optimiser is not defined yet in load_optimiser()")
end
end
function init_neuron!(id::Int64, n::passthrough_neuron, n_params::Dict, kfnParams::Dict)
n.id = id
n.knowledgeFnName = kfnParams[:knowledgeFnName]
end
# function init_neuron!(id::Int64, n::lif_neuron, kfnParams::Dict)
# n.id = id
# n.knowledgeFnName = kfnParams[:knowledgeFnName]
# subscription_options = shuffle!([1:(kfnParams[:input_neuron_number]+kfnParams[:compute_neuron_number])...])
# if typeof(kfnParams[:synaptic_connection_number]) == String
# percent = parse(Int, kfnParams[:synaptic_connection_number][1:end-1]) / 100
# synaptic_connection_number = floor(length(subscription_options) * percent)
# n.subscriptionList = [pop!(subscription_options) for i = 1:synaptic_connection_number]
# end
# filter!(x -> x != n.id, n.subscriptionList)
# n.epsilonRec = zeros(length(n.subscriptionList))
# n.w_rec = Random.rand(length(n.subscriptionList))
# n.wRecChange = zeros(length(n.subscriptionList))
# n.reg_voltage_b = zeros(length(n.subscriptionList))
# n.alpha = calculate_α(n)
# end
function init_neuron!(id::Int64, n::lif_neuron, n_params::Dict, kfnParams::Dict)
n.id = id
n.knowledgeFnName = kfnParams[:knowledgeFnName]
subscription_options = shuffle!([1:kfnParams[:total_neurons]...])
subscription_numbers = Int(floor(n_params[:synaptic_connection_number] *
kfnParams[:total_neurons] / 100.0))
n.subscriptionList = [pop!(subscription_options) for i = 1:subscription_numbers]
# prevent subscription to itself by removing this neuron id
filter!(x -> x != n.id, n.subscriptionList)
n.synapticStrength = normalize!(rand(length(n.subscriptionList)), 1)
n.epsilonRec = zeros(length(n.subscriptionList))
n.w_rec = Random.rand(length(n.subscriptionList))
n.wRecChange = zeros(length(n.subscriptionList))
n.alpha = calculate_α(n)
end
function init_neuron!(id::Int64, n::alif_neuron, n_params::Dict,
kfnParams::Dict)
n.id = id
n.knowledgeFnName = kfnParams[:knowledgeFnName]
subscription_options = shuffle!([1:kfnParams[:total_neurons]...])
subscription_numbers = Int(floor(n_params[:synaptic_connection_number] *
kfnParams[:total_neurons] / 100.0))
n.subscriptionList = [pop!(subscription_options) for i = 1:subscription_numbers]
# prevent subscription to itself by removing this neuron id
filter!(x -> x != n.id, n.subscriptionList)
n.synapticStrength = normalize!(rand(length(n.subscriptionList)), 1)
n.epsilonRec = zeros(length(n.subscriptionList))
n.w_rec = Random.rand(length(n.subscriptionList))
n.wRecChange = zeros(length(n.subscriptionList))
# the more time has passed from the last time neuron was activated, the more
# neuron membrane potential is reduced
n.alpha = calculate_α(n)
n.rho = calculate_ρ(n)
n.epsilonRecA = zeros(length(n.subscriptionList))
end
function init_neuron!(id::Int64, n::linear_neuron, n_params::Dict, kfnParams::Dict)
n.id = id
n.knowledgeFnName = kfnParams[:knowledgeFnName]
subscription_options = shuffle!([kfnParams[:total_input_port]+1 : kfnParams[:total_neurons]...])
subscription_numbers = Int(floor(n_params[:synaptic_connection_number] *
kfnParams[:total_compute_neuron] / 100.0))
n.subscriptionList = [pop!(subscription_options) for i = 1:subscription_numbers]
n.synapticStrength = normalize!(rand(length(n.subscriptionList)), 1)
n.epsilonRec = zeros(length(n.subscriptionList))
n.w_rec = Random.rand(length(n.subscriptionList))
n.wRecChange = zeros(length(n.subscriptionList))
n.alpha = calculate_k(n)
end
""" Make a neuron intended for use with knowledgeFn
"""
function init_neuron(id::Int64, n_params::Dict, kfnParams::Dict)
n = instantiate_custom_types(n_params)
init_neuron!(id, n, n_params, kfnParams)
return n
end
""" This function instantiate Ironpen type.
# Example
new_model = instantiate_custom_types("model")
"""
function instantiate_custom_types(params::Union{Dict,Nothing} = nothing)
type = string(split(params[:type], ".")[end])
if type == "model"
return model()
elseif type == "knowledgeFn"
return knowledgeFn()
elseif type == "passthrough_neuron"
return passthrough_neuron(params)
elseif type == "lif_neuron"
return lif_neuron(params)
elseif type == "alif_neuron"
return alif_neuron(params)
elseif type == "linear_neuron"
return linear_neuron(params)
else
return nothing
end
end
""" Add a new neuron into a knowledgeFn
# Example
add_neuron!(kfn.kfnParams[:lif_neuron_params], kfn)
"""
# function add_neuron!(neuron_Dict::Dict, kfn::knowledgeFn)
# id = length(kfn.neuronsArray) + 1
# neuron = init_neuron(id, neuron_Dict, kfn.kfnParams,
# total_neurons = (length(kfn.neuronsArray) + 1))
# push!(kfn.neuronsArray, neuron)
# # Randomly select an output neuron to add a new neuron to
# add_n_output_n!(Random.rand(kfn.outputNeuronsArray), id)
# end
calculate_α(neuron::lif_neuron) = exp(-neuron.delta / neuron.tau_m)
calculate_α(neuron::alif_neuron) = exp(-neuron.delta / neuron.tau_m)
calculate_ρ(neuron::alif_neuron) = exp(-neuron.delta / neuron.tau_a)
calculate_k(neuron::linear_neuron) = exp(-neuron.delta / neuron.tau_out)
#------------------------------------------------------------------------------------------------100
end # module end