diff --git a/src/pomdp/inference.jl b/src/pomdp/inference.jl
index 23503ef..d0bf041 100644
--- a/src/pomdp/inference.jl
+++ b/src/pomdp/inference.jl
@@ -96,7 +96,7 @@ function fixed_point_iteration(A::Vector{Any}, obs::Vector{Vector{Real}}, num_ob
     likelihood = spm_log_single(likelihood)
 
     # Initialize posterior and prior
-    qs = Array{Any}(undef, n_factors)
+    qs = Vector{Vector{Real}}(undef, n_factors)
     for factor in 1:n_factors
         qs[factor] = ones(Real,num_states[factor]) / num_states[factor]
     end
@@ -113,7 +113,7 @@ function fixed_point_iteration(A::Vector{Any}, obs::Vector{Vector{Real}}, num_ob
     # Single factor condition
     if n_factors == 1
         qL = spm_dot(likelihood, qs[1])  
-        return to_array_of_any(softmax(qL .+ prior[1]))
+        return [softmax(qL .+ prior[1])]
     else
         # Run Iteration 
         curr_iter = 0
diff --git a/src/utils/utils.jl b/src/utils/utils.jl
index 5433a60..ea9d4f0 100644
--- a/src/utils/utils.jl
+++ b/src/utils/utils.jl
@@ -55,19 +55,6 @@ function get_model_dimensions(A = nothing, B = nothing)
 end
 
 
-""" Equivalent to pymdp's "to_obj_array" """
-function to_array_of_any(arr::Array)
-    # Check if arr is already an array of arrays
-    if typeof(arr) == Array{Array,1}
-        return arr
-    end
-    # Create an array_out and assign squeezed array to the first element
-    obj_array_out = Array{Any,1}(undef, 1)
-    obj_array_out[1] = dropdims(arr, dims = tuple(findall(size(arr) .== 1)...))  
-    return obj_array_out
-end
-
-
 """ Selects the highest value from Array -- used for deterministic action sampling """
 function select_highest(options_array::Array{Float64})
     options_with_idx = [(i, option) for (i, option) in enumerate(options_array)]