Efficient and Robust Automated Machine Learning

By Matthias Feurer, Aaron Klein, Katharina Eggensperger, Jost Tobias Springenberg, Manuel Blum, Frank Hutter. 2015

1,941 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Optimization & Training

Citations in the corpus

Train Large, Then Compress: Rethinking Model Size for Efficient Training and Inference of Transformers (2020, methodology): “More generally, our findings may impact the design of automated methods for solving/optimizing machine learning problems (Feurer et al., 2015; Zoph & Le, 2017; Jaderberg et al., 2017).”
Regularized Evolution for Image Classifier Architecture Search (2018, background): “2017; Domhan, Springenberg, and Hutter 2017), warm-starting and ensembling (Feurer et al. 2015), parallelization, reward shaping and early stopping (Zhong, Yan, and Liu 2018) or Net2Net transformations (Cai et al.”
AutoML: A Survey of the State-of-the-Art (2019, neutral): “Although there are several pipeline libraries proposed, such as TPOT [138] and Auto-Sklearn [139], they all lack the procedure of data collection, a process that is usually ﬁnished manually and therefore time-consuming and tedious.”
Hyper-Parameter Optimization: A Review of Algorithms and Applications (2020, background): “To save the rare resource of experts experience, automated machine learning (AutoML) has been proposed as a burgeoning technology to design and train neural networks automatically, at the cost of computational resources (Feurer et al., 2015; Katz et al., 2016; Bello et al., 2017; Zoph and Le, 2016; ”
Towards modular and programmable architecture search (2019, background): “Systems such as Auto-Sklearn [19], TPOT [20], and Auto-Keras [21] have been developed for specific use-cases (e.”
Multi-Fidelity Automatic Hyper-Parameter Tuning via Transfer Series Expansion (2019, methodology): “Further more, some open-source tools (Feurer et al. 2015; Thornton et al. 2013) based on those methods make application of AutoML easier.”
Automated Machine Learning: State-of-The-Art and Open Challenges (2019, background): “Auto-Sklearn(9) has been implemented on top of ScikitLearn(10), a popular Python machine learning package [39].”
Benchmark and Survey of Automated Machine Learning Frameworks (2019, methodology): “Auto-Sklearn auto-sklearn (Feurer et al., 2015a, 2018) is a tool for building classification and regression pipelines.”
Survey on Automated Machine Learning (2019, methodology): “auto-sklearn (Feurer et al., 2015a, 2018) is a tool for building classification and regression pipelines.”
Random Search and Reproducibility for Neural Architecture Search (2019, background): “In traditional hyperparameter optimization, it is standard for empirical results to be reported over 10 independent experimental runs [14, 25, 23, 27].”
Towards Automated Deep Learning: Efficient Joint Neural Architecture and Hyperparameter Search (2018, background): “This process is suboptimal for various reasons: • The resulting approach is not an anytime approach and does not satisfy the requirement of an automated machine learning (AutoML) system (Feurer et al., 2015) to make predictions after a given time budget.”
Auto-Keras: An Efficient Neural Architecture Search System (2018, methodology): “For example, it has been used in hyperparameter tuning for machine learning models [13, 15, 17, 35], in which Bayesian optimization searches among different combinations of hyperparameters.”
Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization (2016, methodology): “(2015a) used 140 binary and multiclass classification datasets from OpenML, but 23 of them are incompatible with the latest version of the OpenML plugin (Feurer, 2015), so we worked with the remaining 117 datasets.”
In AI We Trust? Factors That Influence Trustworthiness of AI-infused Decision-Making Processes (2019, methodology): “compare decisions made entirely by AI to decisions made by a human supported by AI guidance. •Trainer. How was the AI trained? Given the current popularity of automated training methods for AI (e.g., [11, 43]), we consider AI models trained by human data scientists vs. those trained by automated AI.”
Meta-Learning: A Survey (2018, neutral): “This warm-starting approach was also used in later work [46], which is discussed in detail in Chap.”
TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning (2016, background): “In auto-sklearn, [5] imposed a short and fixed pipeline structure of a data preprocessor, a feature preprocessor, and a model.”
Tune: A Research Platform for Distributed Model Selection and Training (2018, methodology): “This implies that in principle, it would be possible to implement Auto-WEKA and Auto-SKLearn on top of Tune, providing distributed execution of these AutoML components.”
A System for Massively Parallel Hyperparameter Tuning (2018, methodology): “, 2017) and AutoSklearn (Feurer et al., 2015) are two established single-machine, single-user systems for hyperparameter optimization.”
Automated Machine Learning: From Principles to Practices (2018, methodology): “In this section, the transformation of search space is mainly divided into four classes: softmax, sparse coding, autoencoder and stochastic relaxation.”