Adding Models for General Use

This guide outlines in detail the specification of the MLJ model interface and provides guidelines for implementing the interface for models intended for general use. For sample implementations, see MLJModels/src.

The machine learning tools provided by MLJ can be applied to the models in any package that imports the package MLJBase and implements the API defined there, as outlined below. For a quick-and-dirty implementation of user-defined models see Simple User Defined Models. To make new models available to all MLJ users, see Where to place code implementing new models.

It is assumed the reader has read Getting Started. To implement the API described here, some familiarity with the following packages is also helpful:

• ScientificTypes.jl (for specifying model requirements of data)

• Distributions.jl (for probabilistic predictions)

• CategoricalArrays.jl (essential if you are implementing a model handling data of Multiclass or OrderedFactor scitype; familiarity with CategoricalPool objects required)

• Tables.jl (if your algorithm needs input data in a novel format).

In MLJ, the basic interface exposed to the user, built atop the model interface described here, is the machine interface. After a first reading of this document, the reader may wish to refer to MLJ Internals for context.

Overview

A model is an object storing hyperparameters associated with some machine learning algorithm. In MLJ, hyperparameters include configuration parameters, like the number of threads, and special instructions, such as "compute feature rankings", which may or may not affect the final learning outcome. However, the logging level (verbosity below) is excluded.

The name of the Julia type associated with a model indicates the associated algorithm (e.g., DecisionTreeClassifier). The outcome of training a learning algorithm is called a fitresult. For ordinary multivariate regression, for example, this would be the coefficients and intercept. For a general supervised model, it is the (generally minimal) information needed to make new predictions.

The ultimate supertype of all models is MLJBase.Model, which has two abstract subtypes:

abstract type Supervised <: Model end
abstract type Unsupervised <: Model end

Supervised models are further divided according to whether they are able to furnish probabilistic predictions of the target (which they will then do by default) or directly predict "point" estimates, for each new input pattern:

abstract type Probabilistic <: Supervised end
abstract type Deterministic <: Supervised end

Further division of model types is realized through Trait declarations.

Associated with every concrete subtype of Model there must be a fit method, which implements the associated algorithm to produce the fitresult. Additionally, every Supervised model has a predict method, while Unsupervised models must have a transform method. More generally, methods such as these, that are dispatched on a model instance and a fitresult (plus other data), are called operations. Probabilistic supervised models optionally implement a predict_mode operation (in the case of classifiers) or a predict_mean and/or predict_median operations (in the case of regressors) although MLJBase also provides fallbacks that will suffice in most cases. Unsupervised models may implement an inverse_transform operation.

New model type declarations and optional clean! method

Here is an example of a concrete supervised model type declaration:

import MLJ

mutable struct RidgeRegressor <: MLJBase.Deterministic
    lambda::Float64
end

Models (which are mutable) should not be given internal constructors. It is recommended that they be given an external lazy keyword constructor of the same name. This constructor defines default values for every field, and optionally corrects invalid field values by calling a clean! method (whose fallback returns an empty message string):

function MLJ.clean!(model::RidgeRegressor)
    warning = ""
    if model.lambda < 0
        warning *= "Need lambda ≥ 0. Resetting lambda=0. "
        model.lambda = 0
    end
    return warning
end

# keyword constructor
function RidgeRegressor(; lambda=0.0)
    model = RidgeRegressor(lambda)
    message = MLJBase.clean!(model)
    isempty(message) || @warn message
    return model
end

Supervised models

The compulsory and optional methods to be implemented for each concrete type SomeSupervisedModel <: MLJBase.Supervised are summarized below. An = indicates the return value for a fallback version of the method.

Summary of methods

Compulsory:

MLJBase.fit(model::SomeSupervisedModel, verbosity::Integer, X, y) -> fitresult, cache, report
MLJBase.predict(model::SomeSupervisedModel, fitresult, Xnew) -> yhat

Optional, to check and correct invalid hyperparameter values:

MLJBase.clean!(model::SomeSupervisedModel) = ""

Optional, to return user-friendly form of fitted parameters:

MLJBase.fitted_params(model::SomeSupervisedModel, fitresult) = fitresult

Optional, to avoid redundant calculations when re-fitting machines associated with a model:

MLJBase.update(model::SomeSupervisedModel, verbosity, old_fitresult, old_cache, X, y) =
   MLJBase.fit(model, verbosity, X, y)

Optional, if SomeSupervisedModel <: Probabilistic:

MLJBase.predict_mode(model::SomeSupervisedModel, fitresult, Xnew) =
    mode.(predict(model, fitresult, Xnew))
MLJBase.predict_mean(model::SomeSupervisedModel, fitresult, Xnew) =
    mean.(predict(model, fitresult, Xnew))
MLJBase.predict_median(model::SomeSupervisedModel, fitresult, Xnew) =
    median.(predict(model, fitresult, Xnew))

Required, if the model is to be registered (findable by general users):

MLJBase.load_path(::Type{<:SomeSupervisedModel})    = ""
MLJBase.package_name(::Type{<:SomeSupervisedModel}) = "Unknown"
MLJBase.package_uuid(::Type{<:SomeSupervisedModel}) = "Unknown"

Strongly recommended, to constrain the form of input data passed to fit and predict:

MLJBase.input_scitype(::Type{<:SomeSupervisedModel}) = Unknown

Strongly recommended, to constrain the form of target data passed to fit:

MLJBase.target_scitype(::Type{<:SomeSupervisedModel}) = Unknown

Optional but recommended:

MLJBase.package_url(::Type{<:SomeSupervisedModel})  = "unknown"
MLJBase.is_pure_julia(::Type{<:SomeSupervisedModel}) = false
MLJBase.package_license(::Type{<:SomeSupervisedModel}) = "unknown"

The form of data for fitting and predicting

The model implementer does not have absolute control over the types of data X, y and Xnew appearing in the fit and predict methods they must implement. Rather, they can specify the scientific type of this data by making appropriate declarations of the traits input_scitype and target_scitype discussed later under Trait declarations.

Important Note. Unless it genuinely makes little sense to do so, the MLJ recommendation is to specify a Table scientific type for X (and hence Xnew) and an AbstractVector scientific type (e.g., AbstractVector{Continuous}) for targets y. Algorithms requiring matrix input can coerce their inputs appropriately; see below.

Additional type coercions

If the core algorithm being wrapped requires data in a different or more specific form, then fit will need to coerce the table into the form desired (and the same coercions applied to X will have to be repeated for Xnew in predict). To assist with common cases, MLJ provides the convenience method MLJBase.matrix. MLJBase.matrix(Xtable) has type Matrix{T} where T is the tightest common type of elements of Xtable, and Xtable is any table.

Other auxiliary methods provided by MLJBase for handling tabular data are: selectrows, selectcols, select and schema (for extracting the size, names and eltypes of a table's columns). See Convenience methods below for details.

Important convention

It is to be understood that the columns of the table X correspond to features and the rows to observations. So, for example, the predict method for a linear regression model might look like predict(model, w, Xnew) = MLJBase.matrix(Xnew)*w, where w is the vector of learned coefficients.

The fit method

A compulsory fit method returns three objects:

MLJBase.fit(model::SomeSupervisedModel, verbosity::Int, X, y) -> fitresult, cache, report

Note. The Int typing of verbosity cannot be omitted.

1. fitresult is the fitresult in the sense above (which becomes an argument for predict discussed below).

2. report is a (possibly empty) NamedTuple, for example, report=(deviance=..., dof_residual=..., stderror=..., vcov=...). Any training-related statistics, such as internal estimates of the generalization error, and feature rankings, should be returned in the report tuple. How, or if, these are generated should be controlled by hyperparameters (the fields of model). Fitted parameters, such as the coefficients of a linear model, do not go in the report as they will be extractable from fitresult (and accessible to MLJ through the fitted_params method described below).

3. The value of cache can be nothing, unless one is also defining an update method (see below). The Julia type of cache is not presently restricted.

It is not necessary for fit to provide type or dimension checks on X or y or to call clean! on the model; MLJ will carry out such checks.

The method fit should never alter hyperparameter values, the sole exception being fields of type <:AbstractRNG. If the package is able to suggest better hyperparameters, as a byproduct of training, return these in the report field.

The verbosity level (0 for silent) is for passing to learning algorithm itself. A fit method wrapping such an algorithm should generally avoid doing any of its own logging.

The fitted_params method

A fitted_params method may be optionally overloaded. It's purpose is to provide MLJ access to a user-friendly representation of the learned parameters of the model (as opposed to the hyperparameters). They must be extractable from fitresult.

MLJBase.fitted_params(model::SomeSupervisedModel, fitresult) -> friendly_fitresult::NamedTuple

For a linear model, for example, one might declare something like friendly_fitresult=(coefs=[...], bias=...).

The fallback is to return (fitresult=fitresult,).

The predict method

A compulsory predict method has the form

MLJBase.predict(model::SomeSupervisedModel, fitresult, Xnew) -> yhat

Here Xnew will have the same form as the X passed to fit.

Prediction types for deterministic responses.

In the case of Deterministic models, yhat should have the same scitype as the y passed to fit (see above). Any CategoricalValue or CategoricalString elements of yhat must have a pool == to the pool of the target y presented in training, even if not all levels appear in the training data or prediction itself. For example, in the case of a univariate target, such as scitype(y) <: AbstractVector{Multiclass{3}}, one requires MLJ.classes(yhat[i]) == MLJ.classes(y[j]) for all admissible i and j. (The method classes is described under Convenience methods below).

Unfortunately, code not written with the preservation of categorical levels in mind poses special problems. To help with this, MLJBase provides three utility methods: int (for converting a CategoricalValue or CategoricalString into an integer, the ordering of these integers being consistent with that of the pool), decoder (for constructing a callable object that decodes the integers back into CategoricalValue/CategoricalString objects), and classes, for extracting all the CategoricalValue or CategoricalString objects sharing the pool of a particular value/string. Refer to Convenience methods below for important details.

Note that a decoder created during fit may need to be bundled with fitresult to make it available to predict during re-encoding. So, for example, if the core algorithm being wrapped by fit expects a nominal target yint of type Vector{<:Integer} then a fit method may look something like this:

function MLJBase.fit(model::SomeSupervisedModel, verbosity, X, y)
    yint = MLJBase.int(y)
    a_target_element = y[1]                    # a CategoricalValue/String
    decode = MLJBase.decoder(a_target_element) # can be called on integers

    core_fitresult = SomePackage.fit(X, yint, verbosity=verbosity)

    fitresult = (decode, core_fitresult)
    cache = nothing
    report = nothing
    return fitresult, cache, report
end

while a corresponding deterministic predict operation might look like this:

function MLJBase.predict(model::SomeSupervisedModel, fitresult, Xnew)
    decode, core_fitresult = fitresult
    yhat = SomePackage.predict(core_fitresult, Xnew)
    return decode.(yhat)  # or decode(yhat) also works
end

For a concrete example, refer to the code for SVMClassifier.

Of course, if you are coding a learning algorithm from scratch, rather than wrapping an existing one, these extra measures may be unnecessary.

Prediction types for probabilistic responses

In the case of Probabilistic models with univariate targets, yhat must be an AbstractVector whose elements are distributions (one distribution per row of Xnew).

Presently, a distribution is any object d for which MLJBase.isdistribution(::d) = true, which is currently restricted to objects subtyping Distributions.Sampleable from the package Distributions.jl.

Use the distribution MLJBase.UnivariateFinite for Probabilistic models predicting a target with Finite scitype (classifiers). In this case each element of the training target y is a CategoricalValue or CategoricalString, as in this contrived example:

using CategoricalArrays
y = Any[categorical([:yes, :no, :no, :maybe, :maybe])...]

Note that, as in this case, we cannot assume y is a CategoricalVector, and we rely on elements for pool information (if we need it); this is accessible using the convenience method MLJ.classes:

julia> yes = y[1]
julia> levels = MLJBase.classes(yes)
3-element Array{CategoricalValue{Symbol,UInt32},1}:
 :maybe
 :no
 :yes

Now supposing that, for some new input pattern, the elements yes = y[1] and no = y[2] are to be assigned respective probabilities of 0.2 and 0.8. Then the corresponding distribution d is constructed as follows:

julia> d = MLJBase.UnivariateFinite([yes, no], [0.2, 0.8])
UnivariateFinite(:yes=>0.2, :maybe=>0.0, :no=>0.8)

julia> pdf(d, yes)
0.2

julia> maybe = y[4]; pdf(d, maybe)
0.0

Alternatively, a dictionary can be passed to the constructor.

See BinaryClassifier for an example of a Probabilistic classifier implementation.

UnivariateFinite(classes, p)

CategoricalElement = Union{CategoricalValue,CategoricalString}

UnivariateFinite(prob_given_class)

classes(d::UnivariateFinite)

Distributions.support(d::UnivariateFinite)

v = categorical(["yes", "maybe", "no", "yes"])
d = UnivariateFinite(v[1:2], [0.3, 0.7])
pdf(d, "yes")     # 0.3
pdf(d, v[1])      # 0.3
pdf(d, "no")      # 0.0
pdf(d, "house")   # throws error
classes(d) # Array{CategoricalString{UInt32}, 1}["maybe", "no", "yes"]
support(d) # Array{CategoricalString{UInt32}, 1}["maybe", "no"]
mode(d)    # CategoricalString{UInt32} "maybe"
rand(d, 5) # Array{CategoricalString{UInt32}, 1}["maybe", "no", "maybe", "maybe", "no"] or similar
d = fit(UnivariateFinite, v)
pdf(d, "maybe") # 0.25 

Important note on binary classifiers. ScientificTypes.jl has no "Binary" scitype distinct from Multiclass{2} or OrderedFactor{2}; Binary is just an alias for Union{Multiclass{2},OrderedFactor{2}}. The target_scitype of a binary classifier will be AbstractVector{<:Binary} and according to the mlj scitype convention, elements of y have type CategoricalValue or CategoricalString, and not Bool. See BinaryClassifier for an example.

Trait declarations

Two trait functions allow the implementer to restrict the types of data X, y and Xnew discussed above. The MLJ task interface uses these traits for data type checks but also for model search. If they are omitted (and your model is registered) then a general user may attempt to use your model with inappropriately typed data.

The trait functions input_scitype and target_scitype take scientific data types as values. We assume here familiarity with ScientificTypes.jl (see Getting Started for the basics).

For example, to ensure that the X presented to the DecisionTreeClassifier fit method is a table whose columns all have Continuous element type (and hence AbstractFloat machine type), one declares

MLJBase.input_scitype(::Type{<:DecisionTreeClassifier}) = MLJBase.Table(MLJBase.Continuous)

or, equivalently,

using ScientificTypes
MLJBase.input_scitype(::Type{<:DecisionTreeClassifier}) = Table(Continuous)

If, instead, columns were allowed to have either: (i) a mixture of Continuous and Missing values, or (ii) Count (i.e., integer) values, then the declaration would be

MLJBase.input_scitype(::Type{<:DecisionTreeClassifier}) = Table(Union{Continuous,Missing},Count)

Similarly, to ensure the target is an AbstractVector whose elements have Finite scitype (and hence CategoricalValue or CategoricalString machine type) we declare

MLJBase.target_scitype(::Type{<:DecisionTreeClassifier}) = AbstractVector{<:Finite}

Multivariate targets

The above remarks continue to hold unchanged for the case multivariate targets. For example, if we declare

target_scitype(SomeSupervisedModel) = AbstractVector{<:Tuple{Continuous,Count}}

then each element of y will be a tuple of type Tuple{AbstractFloat,Integer}. For predicting variable length sequences of, say, binary values (CategoricalValues or CategoricalStrings with some common size-two pool) we declare

target_scitype(SomeSupervisedModel) = AbstractVector{<:NTuple{<:Binary}}

The trait functions controlling the form of data are summarized as follows:

Additional trait functions tell MLJ's @load macro how to find your model if it is registered, and provide other self-explanatory metadata about the model:

Here is the complete list of trait function declarations for DecisionTreeClassifier (source):

MLJBase.input_scitype(::Type{<:DecisionTreeClassifier}) = MLJBase.Table(MLJBase.Continuous)
MLJBase.target_scitype(::Type{<:DecisionTreeClassifier}) = AbstractVector{<:MLJBase.Finite}
MLJBase.load_path(::Type{<:DecisionTreeClassifier}) = "MLJModels.DecisionTree_.DecisionTreeClassifier"
MLJBase.package_name(::Type{<:DecisionTreeClassifier}) = "DecisionTree"
MLJBase.package_uuid(::Type{<:DecisionTreeClassifier}) = "7806a523-6efd-50cb-b5f6-3fa6f1930dbb"
MLJBase.package_url(::Type{<:DecisionTreeClassifier}) = "https://github.com/bensadeghi/DecisionTree.jl"
MLJBase.is_pure_julia(::Type{<:DecisionTreeClassifier}) = true

You can test all your declarations of traits by calling MLJBase.info_dict(SomeModel).

Iterative models and the update! method

An update method may be optionally overloaded to enable a call by MLJ to retrain a model (on the same training data) to avoid repeating computations unnecessarily.

MLJBase.update(model::SomeSupervisedModel, verbosity, old_fitresult, old_cache, X, y) -> fitresult, cache, report

If an MLJ Machine is being fit! and it is not the first time, then update is called instead of fit, unless the machine fit! has been called with a new rows keyword argument. However, MLJBase defines a fallback for update which just calls fit. For context, see MLJ Internals.

Learning networks wrapped as models constitute one use-case (see Composing Models): one would like each component model to be retrained only when hyperparameter changes "upstream" make this necessary. In this case MLJ provides a fallback (specifically, the fallback is for any subtype of SupervisedNetwork = Union{DeterministicNetwork,ProbabilisticNetwork}). A second more generally relevant use-case is iterative models, where calls to increase the number of iterations only restarts the iterative procedure if other hyperparameters have also changed. For an example, see the MLJ ensemble code.

A third use-case is to avoid repeating time-consuming preprocessing of X and y required by some models.

In the event that the argument fitresult (returned by a preceding call to fit) is not sufficient for performing an update, the author can arrange for fit to output in its cache return value any additional information required (for example, pre-processed versions of X and y), as this is also passed as an argument to the update method.

Unsupervised models

TODO

This is basically the same but with no target y appearing in the signatures, and no target_scitype trait to declare.

Convenience methods

int(X::CategoricalArray)
int(W::Array{<:CategoricalString})
int(W::Array{<:CategoricalValue})

julia> v = categorical([:c, :b, :c, :a])
julia> levels(v)
3-element Array{Symbol,1}:
 :a
 :b
 :c
julia> int(v)
4-element Array{UInt32,1}:
 0x00000003
 0x00000002
 0x00000003
 0x00000001

classes(x)

julia> v = categorical([:c, :b, :c, :a])
julia> levels(v)
3-element Array{Symbol,1}:
 :a
 :b
 :c
julia> classes(v[4])
3-element Array{CategoricalValue{Symbol,UInt32},1}:
 :a
 :b
 :c

d = decoder(x)

julia> v = categorical([:c, :b, :c, :a])
julia> int(v)
4-element Array{UInt32,1}:
 0x00000003
 0x00000002
 0x00000003
 0x00000001
julia> d = decoder(v[3])
julia> d(int(v)) == v
true

MLJBase.matrix(X; transpose=false)

MLJBase.table(columntable; prototype=nothing)

MLJBase.table(A::AbstractMatrix; names=nothing, prototype=nothing)

select(X, r, c)

selectrows(X, r)

selectrows(X, r)

selectcols(X, c)

schema(X)

julia> X = (ncalls=[1, 2, 4], mean_delay=[2.0, 5.7, 6.0])
julia> schema(X)
(names = (:ncalls, :mean_delay), 
 types = (Int64, Float64), 
 scitypes = (Count, Continuous))

nrows(X)

scitype(x)

scitype_union(A)

Where to place code implementing new models

Note that different packages can implement models having the same name without causing conflicts, although an MLJ user cannot simultaneously load two such models.

There are two options for making a new model implementation available to all MLJ users:

1. Native implementations (preferred option). The implementation code lives in the same package that contains the learning algorithms implementing the interface. In this case, it is sufficient to open an issue at MLJ requesting the package to be registered with MLJ. Registering a package allows the MLJ user to access its models' metadata and to selectively load them.

2. External implementations (short-term alternative). The model implementation code is necessarily separate from the package SomePkg defining the learning algorithm being wrapped. In this case, the recommended procedure is to include the implementation code at MLJModels/src via a pull-request, and test code at MLJModels/test. Assuming SomePkg is the only package imported by the implementation code, one needs to: (i) register SomePkg with MLJ as explained above; and (ii) add a corresponding @require line in the PR to MLJModels/src/MLJModels.jl to enable lazy-loading of that package by MLJ (following the pattern of existing additions). If other packages must be imported, add them to the MLJModels project file after checking they are not already there. If it is really necessary, packages can be also added to Project.toml for testing purposes.

Additionally, one needs to ensure that the implementation code defines the package_name and load_path model traits appropriately, so that MLJ's @load macro can find the necessary code (see MLJModels/src for examples). The @load command can only be tested after registration. If changes are made, lodge an new issue at MLJ requesting your changes to be updated.