Argmax

struct ArgmaxBenchmark{E} <: AbstractBenchmark

Basic benchmark problem with an argmax as the CO algorithm.

Fields

• instance_dim::Int64: instances dimension, total number of classes

• nb_features::Int64: number of features

• encoder::Any: true mapping between features and costs

ArgmaxBenchmark(
;
    instance_dim,
    nb_features,
    seed
) -> ArgmaxBenchmark{E} where E<:(Flux.Chain{T} where T<:Tuple{Flux.Dense{typeof(identity), Matrix{Float32}}, typeof(vec)})

Custom constructor for ArgmaxBenchmark.

one_hot_argmax(
    z::AbstractArray{R<:Real, 1};
    kwargs...
) -> Any

One-hot encoding of the argmax function.

generate_dataset(bench::ArgmaxBenchmark; ...) -> Any
generate_dataset(
    bench::ArgmaxBenchmark,
    dataset_size::Int64;
    seed,
    noise_std
) -> Any

Generate a dataset of labeled instances for the argmax problem.

generate_maximizer(
    bench::ArgmaxBenchmark
) -> typeof(DecisionFocusedLearningBenchmarks.Argmax.one_hot_argmax)

Return an argmax maximizer.

generate_statistical_model(
    bench::ArgmaxBenchmark;
    seed
) -> Flux.Chain{T} where T<:Tuple{Flux.Dense{typeof(identity), Matrix{Float32}}, typeof(vec)}

Initialize a linear model for bench using Flux.