A Framework for Model Search Across Multiple Machine Learning Implementations

Several recently devised machine learning (ML) algorithms have shown improved accuracy for various predictive problems. Model searches, which explore to find an optimal ML algorithm and hyperparameter values for the target problem, play a critical role in such improvements. During a model search, data scientists typically use multiple ML implementations to construct several predictive models; however, it takes significant time and effort to employ multiple ML implementations due to the need to learn how to use them, prepare input data in several different formats, and compare their outputs. Our proposed framework addresses these issues by providing simple and unified coding method. It has been designed with the following two attractive features: i) new machine learning implementations can be added easily via common interfaces between the framework and ML implementations and ii) it can be scaled to handle large model configuration search spaces via profile-based scheduling. The results of our evaluation indicate that, with our framework, implementers need only write 55-144 lines of code to add a new ML implementation. They also show that ours was the fastest framework for the HIGGS dataset, and the second-fastest for the SECOM dataset.

[1]  Ramesh Raskar,et al.  Designing Neural Network Architectures using Reinforcement Learning , 2016, ICLR.

[2]  Tim Kraska,et al.  MLbase: A Distributed Machine-learning System , 2013, CIDR.

[3]  David D. Cox,et al.  Hyperopt: A Python Library for Optimizing the Hyperparameters of Machine Learning Algorithms , 2013, SciPy.

[4]  Scott Shenker,et al.  Spark: Cluster Computing with Working Sets , 2010, HotCloud.

[5]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[6]  Ravi Sethi,et al.  The Complexity of Flowshop and Jobshop Scheduling , 1976, Math. Oper. Res..

[7]  A. Asuncion,et al.  UCI Machine Learning Repository, University of California, Irvine, School of Information and Computer Sciences , 2007 .

[8]  Ameet Talwalkar,et al.  MLlib: Machine Learning in Apache Spark , 2015, J. Mach. Learn. Res..

[9]  Jasper Snoek,et al.  Practical Bayesian Optimization of Machine Learning Algorithms , 2012, NIPS.

[10]  Tim Kraska,et al.  Automating model search for large scale machine learning , 2015, SoCC.

[11]  Olatunji Ruwase,et al.  HyperDrive: exploring hyperparameters with POP scheduling , 2017, Middleware.

[12]  Zhi-Hua Zhou,et al.  Experienced Optimization with Reusable Directional Model for Hyper-Parameter Search , 2018, IJCAI.

[13]  Donald R. Jones,et al.  Efficient Global Optimization of Expensive Black-Box Functions , 1998, J. Glob. Optim..

[14]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[15]  D. Wolpert,et al.  No Free Lunch Theorems for Search , 1995 .

[16]  Kevin Leyton-Brown,et al.  Auto-WEKA: combined selection and hyperparameter optimization of classification algorithms , 2012, KDD.

[17]  Tianqi Chen,et al.  XGBoost: A Scalable Tree Boosting System , 2016, KDD.

[18]  Yoshua Bengio,et al.  Random Search for Hyper-Parameter Optimization , 2012, J. Mach. Learn. Res..

[19]  Kazuyuki Shudo,et al.  A Framework for Searching a Predictive Model , 2018 .

[20]  Christopher Kiekintveld,et al.  Initial Exploration of Machine Learning to Predict Customer Demand in an Energy Market Simulation , 2013, AAAI Workshop: Trading Agent Design and Analysis.

[21]  T.B. Trafalis,et al.  Kernel principal component analysis and support vector machines for stock price prediction , 2004, 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541).

[22]  Martin Wattenberg,et al.  Ad click prediction: a view from the trenches , 2013, KDD.

[23]  Natalia V. Shakhlevich,et al.  NP-hardness of Shop-scheduling Problems with Three Jobs , 1995, Discret. Appl. Math..