[
https://issues.apache.org/jira/browse/SPARK-24359?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hossein Falaki updated SPARK-24359:
-----------------------------------
Attachment: SparkML_ ML Pipelines in R-v3.pdf
> SPIP: ML Pipelines in R
> -----------------------
>
> Key: SPARK-24359
> URL: https://issues.apache.org/jira/browse/SPARK-24359
> Project: Spark
> Issue Type: Improvement
> Components: SparkR
> Affects Versions: 3.0.0
> Reporter: Hossein Falaki
> Priority: Major
> Labels: SPIP
> Attachments: SparkML_ ML Pipelines in R-v2.pdf, SparkML_ ML Pipelines
> in R-v3.pdf, SparkML_ ML Pipelines in R.pdf
>
>
> h1. Background and motivation
> SparkR supports calling MLlib functionality with an [R-friendly
> API|https://docs.google.com/document/d/10NZNSEurN2EdWM31uFYsgayIPfCFHiuIu3pCWrUmP_c/].
> Since Spark 1.5 the (new) SparkML API which is based on [pipelines and
> parameters|https://docs.google.com/document/d/1rVwXRjWKfIb-7PI6b86ipytwbUH7irSNLF1_6dLmh8o]
> has matured significantly. It allows users build and maintain complicated
> machine learning pipelines. A lot of this functionality is difficult to
> expose using the simple formula-based API in SparkR.
> We propose a new R package, _SparkML_, to be distributed along with SparkR as
> part of Apache Spark. This new package will be built on top of SparkR’s APIs
> to expose SparkML’s pipeline APIs and functionality.
> *Why not SparkR?*
> SparkR package contains ~300 functions. Many of these shadow functions in
> base and other popular CRAN packages. We think adding more functions to
> SparkR will degrade usability and make maintenance harder.
> *Why not sparklyr?*
> sparklyr is an R package developed by RStudio Inc. to expose Spark API to R
> users. sparklyr includes MLlib API wrappers, but to the best of our knowledge
> they are not comprehensive. Also we propose a code-gen approach for this
> package to minimize work needed to expose future MLlib API, but sparklyr’s
> API is manually written.
> h1. Target Personas
> * Existing SparkR users who need more flexible SparkML API
> * R users (data scientists, statisticians) who wish to build Spark ML
> pipelines in R
> h1. Goals
> * R users can install SparkML from CRAN
> * R users will be able to import SparkML independent from SparkR
> * After setting up a Spark session R users can
> ** create a pipeline by chaining individual components and specifying their
> parameters
> ** tune a pipeline in parallel, taking advantage of Spark
> ** inspect a pipeline’s parameters and evaluation metrics
> ** repeatedly apply a pipeline
> * MLlib contributors can easily add R wrappers for new MLlib Estimators and
> Transformers
> h1. Non-Goals
> * Adding new algorithms to SparkML R package which do not exist in Scala
> * Parallelizing existing CRAN packages
> * Changing existing SparkR ML wrapping API
> h1. Proposed API Changes
> h2. Design goals
> When encountering trade-offs in API, we will chose based on the following
> list of priorities. The API choice that addresses a higher priority goal will
> be chosen.
> # *Comprehensive coverage of MLlib API:* Design choices that make R coverage
> of future ML algorithms difficult will be ruled out.
> * *Semantic clarity*: We attempt to minimize confusion with other packages.
> Between consciousness and clarity, we will choose clarity.
> * *Maintainability and testability:* API choices that require manual
> maintenance or make testing difficult should be avoided.
> * *Interoperability with rest of Spark components:* We will keep the R API
> as thin as possible and keep all functionality implementation in JVM/Scala.
> * *Being natural to R users:* Ultimate users of this package are R users and
> they should find it easy and natural to use.
> The API will follow familiar R function syntax, where the object is passed as
> the first argument of the method: do_something(obj, arg1, arg2). All
> functions are snake_case (e.g., {{spark_logistic_regression()}} and
> {{set_max_iter()}}). If a constructor gets arguments, they will be named
> arguments. For example:
> {code:java}
> > lr <- set_reg_param(set_max_iter(spark.logistic.regression()), 10),
> > 0.1){code}
> When calls need to be chained, like above example, syntax can nicely
> translate to a natural pipeline style with help from very popular[ magrittr
> package|https://cran.r-project.org/web/packages/magrittr/index.html]. For
> example:
> {code:java}
> > logistic_regression() %>% set_max_iter(10) %>% set_reg_param(0.01) ->
> > lr{code}
> h2. Namespace
> All new API will be under a new CRAN package, named SparkML. The package
> should be usable without needing SparkR in the namespace. The package will
> introduce a number of S4 classes that inherit from four basic classes. Here
> we will list the basic types with a few examples. An object of any child
> class can be instantiated with a function call that starts with {{spark_}}.
> h2. Pipeline & PipelineStage
> A pipeline object contains one or more stages.
> {code:java}
> > pipeline <- spark_pipeline() %>% set_stages(stage1, stage2, stage3){code}
> Where stage1, stage2, etc are S4 objects of a PipelineStage and pipeline is
> an object of type Pipeline.
> h2. Transformers
> A Transformer is an algorithm that can transform one SparkDataFrame into
> another SparkDataFrame.
> *Example API:*
> {code:java}
> > tokenizer <- spark_tokenizer() %>%
> set_input_col(“text”) %>%
> set_output_col(“words”)
> > tokenized.df <- tokenizer %>% transform(df) {code}
> h2. Estimators
> An Estimator is an algorithm which can be fit on a SparkDataFrame to produce
> a Transformer. E.g., a learning algorithm is an Estimator which trains on a
> DataFrame and produces a model.
> *Example API:*
> {code:java}
> lr <- spark_logistic_regression() %>%
> set_max_iter(10) %>%
> set_reg_param(0.001){code}
> h2. Evaluators
> An evaluator computes metrics from predictions (model outputs) and returns a
> scalar metric.
> *Example API:*
> {code:java}
> lr.eval <- spark_regression_evaluator(){code}
> h2. Miscellaneous Classes
> MLlib pipelines have a variety of miscellaneous classes that serve as helpers
> and utilities. For example an object of ParamGridBuilder is used to build a
> grid search pipeline. Another example is ClusteringSummary.
> *Example API:*
> {code:java}
> > grid <- param_grid_builder() %>%
> add_grid(reg_param(lr), c(0.1, 0.01)) %>%
> add_grid(fit_intercept(lr), c(TRUE, FALSE)) %>%
> add_grid(elastic_net_param(lr), c(0.0, 0.5, 1.0))
> > model <- train_validation_split() %>%
> set_estimator(lr) %>%
> set_evaluator(spark_regression_evaluator()) %>%
> set_estimator_param_maps(grid) %>%
> set_train_ratio(0.8) %>%
> set_parallelism(2) %>%
> fit(){code}
> Pipeline Persistence
> SparkML package will fix a longstanding issue with SparkR model persistence
> SPARK-15572. SparkML will directly wrap MLlib pipeline persistence API.
> *API example:*
> {code:java}
> > model <- pipeline %>% fit(training)
> > model %>% spark_write_pipeline(overwrite = TRUE, path = “...”){code}
> h1. Design Sketch
> We propose using code generation from Scala to produce comprehensive API
> wrappers in R. For more details please see the attached design document.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
