Efficient AutoML Pipeline Search with Matrix and Tensor Factorization

Abstract

Data scientists seeking a good supervised learning model on a new dataset have many choices to make: they must preprocess the data, select features, possibly reduce the dimension, select an estimation algorithm, and choose hyperparameters for each of these pipeline components. With new pipeline components comes a combinatorial explosion in the number of choices! In this work, we design a new AutoML system to address this challenge: an automated system to design a supervised learning pipeline. Our system uses matrix and tensor factorization as surrogate models to model the combinatorial pipeline search space. Under these models, we develop greedy experiment design protocols to efficiently gather information about a new dataset. Experiments on large corpora of real-world classification problems demonstrate the effectiveness of our approach.

arXiv commentThis is an extended version of AutoML Pipeline Selection: Efficiently Navigating the Combinatorial Space (DOI: 10.1145/3394486.3403197) at KDD 2020

References in corpus (11)

• Scikit-learn: Machine Learning in Python
• Practical Bayesian Optimization of Machine Learning Algorithms
• Theoretical Models of Learning to Learn
• OpenML: networked science in machine learning
• Meta-Learning: A Survey
• OBOE: Collaborative Filtering for AutoML Model Selection
• Probabilistic Matrix Factorization for Automated Machine Learning
• AlphaD3M: Machine Learning Pipeline Synthesis
• OpenML-Python: an extensible Python API for OpenML
• AutoML using Metadata Language Embeddings
• Polynomial Matrix Completion for Missing Data Imputation and Transductive Learning