cavis/arbiter
Alex Black a25bb6a11c
Unit/integration test split + test speedup (#166)
* Add maven profile + base tests methods for integration tests

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Switch from system property to environment variable; seems more reliable in intellij

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Add nd4j-common-tests module, and common base test; cleanup

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Ensure all ND4J tests extend BaseND4JTest

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Test spam reduction, import fix

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Add test logging to nd4j-aeron

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fix unintended change

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Reduce sprint test log spam

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* More test spam cleanup

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Significantly speed up TSNE tests

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* W2V iterator test unit/integration split

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* More NLP test speedups

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Avoid debug/verbose mode leaking between tests

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* test tweak

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Arbiter extends base DL4J test

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Arbiter test speedup

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* nlp-uima test speedup

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* More test speedups

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fix ND4J base test

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Few small ND4J test speed improvements

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* DL4J tests speedup

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* More tweaks

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Even more test speedups

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* More tweaks

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Various test fixes

Signed-off-by: Alex Black <blacka101@gmail.com>

* More test fixes

Signed-off-by: Alex Black <blacka101@gmail.com>

* Add ability to specify number of threads for C++ ops in BaseDL4JTest and BaseND4JTest

Signed-off-by: Alex Black <blacka101@gmail.com>

* nd4j-aeron test profile fix for CUDA

Signed-off-by: Alex Black <blacka101@gmail.com>
2020-01-22 22:27:01 +11:00
..
arbiter-core Unit/integration test split + test speedup (#166) 2020-01-22 22:27:01 +11:00
arbiter-deeplearning4j Unit/integration test split + test speedup (#166) 2020-01-22 22:27:01 +11:00
arbiter-server Unit/integration test split + test speedup (#166) 2020-01-22 22:27:01 +11:00
arbiter-ui Add support for CUDA 10.2 (#89) 2019-11-29 16:31:03 +11:00
ci Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
contrib Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
.travis.yml Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
README.md Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
buildmultiplescalaversions.sh Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
pom.xml Add support for CUDA 10.2 (#89) 2019-11-29 16:31:03 +11:00

README.md

Arbiter

A tool dedicated to tuning (hyperparameter optimization) of machine learning models. Part of the DL4J Suite of Machine Learning / Deep Learning tools for the enterprise.

Modules

Arbiter contains the following modules:

  • arbiter-core: Defines the API and core functionality, and also contains functionality for the Arbiter UI
  • arbiter-deeplearning4j: For hyperparameter optimization of DL4J models (MultiLayerNetwork and ComputationGraph networks)

Hyperparameter Optimization Functionality

The open-source version of Arbiter currently defines two methods of hyperparameter optimization:

  • Grid search
  • Random search

For optimization of complex models such as neural networks (those with more than a few hyperparameters), random search is superior to grid search, though Bayesian hyperparameter optimization schemes For a comparison of random and grid search methods, see Random Search for Hyper-parameter Optimization (Bergstra and Bengio, 2012).

Core Concepts and Classes in Arbiter for Hyperparameter Optimization

In order to conduct hyperparameter optimization in Arbiter, it is necessary for the user to understand and define the following:

  • Parameter Space: A ParameterSpace<P> specifies the type and allowable values of hyperparameters for a model configuration of type P. For example, P could be a MultiLayerConfiguration for DL4J
  • Candidate Generator: A CandidateGenerator<C> is used to generate candidate models configurations of some type C. The following implementations are defined in arbiter-core:
    • RandomSearchCandidateGenerator
    • GridSearchCandidateGenerator
  • Score Function: A ScoreFunction<M,D> is used to score a model of type M given data of type D. For example, in DL4J a score function might be used to calculate the classification accuracy from a DataSetIterator
    • A key concept here is that they score is a single numerical (double precision) value that we either want to minimize or maximize - this is the goal of hyperparameter optimization
  • Termination Conditions: One or more TerminationCondition instances must be provided to the OptimizationConfiguration. TerminationCondition instances are used to control when hyperparameter optimization should be stopped. Some built-in termination conditions:
    • MaxCandidatesCondition: Terminate if more than the specified number of candidate hyperparameter configurations have been executed
    • MaxTimeCondition: Terminate after a specified amount of time has elapsed since starting the optimization
  • Result Saver: The ResultSaver<C,M,A> interface is used to specify how the results of each hyperparameter optimization run should be saved. For example, whether saving should be done to local disk, to a database, to HDFS, or simply stored in memory.
    • Note that ResultSaver.saveModel method returns a ResultReference object, which provides a mechanism for re-loading both the model and score from wherever it may be saved.
  • Optimization Configuration: An OptimizationConfiguration<C,M,D,A> ties together the above configuration options in a fluent (builder) pattern.
  • Candidate Executor: The CandidateExecutor<C,M,D,A> interface provides a layer of abstraction between the configuration and execution of each instance of learning. Currently, the only option is the LocalCandidateExecutor, which is used to execute learning on a single machine (in the current JVM). In principle, other execution methods (for example, on Spark or cloud computing machines) could be implemented.
  • Optimization Runner: The OptimizationRunner uses an OptimizationConfiguration and a CandidateExecutor to actually run the optimization, and save the results.

Optimization of DeepLearning4J Models

(This section: forthcoming)