Identifying Key Algorithm Parameters and Instance Features Using Forward Selection

By Frank Hutter, Holger H. Hoos, Kevin Leyton-Brown. 2013

80 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications, Signal Processing

How researchers see this paper

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

In Artificial Intelligence

In Computer Networks and Communications

In Signal Processing

Citations in the corpus

Taking the Human Out of the Loop: A Review of Bayesian Optimization (2016, background): “Many researchers have noted that for certain classes of problems most dimensions do not change the objective function significantly; examples include hyperparameter optimization for neural networks and deep belief networks [17] and automatic configuration of state-of-the-art algorithms for solving N”
An Efficient Approach for Assessing Hyperparameter Importance (2014, background): “…optimization has targeted the case where most hyperparameters are truly unimportant (Chen et al., 2012; Wang et al., 2013), and several applications have yielded evidence that some hyperparameters indeed tend to be much more important than others (Bergstra and Bengio, 2012; Hutter et al., 2013).”
Beyond Manual Tuning of Hyperparameters (2015, background): “Other initiatives aiming to explain performance include feature selection of important hyperparameters [33] and ablation analysis [19].”
UvA-DARE (Digital Academic Repository) Bayesian Optimization in High Dimensions via Random Embeddings (2013, unknown): “tion of state-of-the-art algorithms for solvingNP-hard problems [Hutter et al., 2013].”
Algorithm runtime prediction: Methods & evaluation (2012, background): “…that suﬃce to yield high-quality predictions, ﬁnding that often as little as ﬁve to ten features yielded predictions as good as the full feature set Hutter et al. [57] extended that work to predictions in the joint space of instance features and algorithm parameters, using arbitrary models.”
Tunability: Importance of Hyperparameters of Machine Learning Algorithms (2018, methodology): “In the following we use random forest as surrogate model because it performs reasonably well and is already an established algorithm for surrogate models in the literature (Eggensperger et al., 2014; Hutter et al., 2013).”
Hyperparameter Importance Across Datasets (2017, background): “Forward selection [18] is based on this principle.”
Efficient Parameter Importance Analysis via Ablation with Surrogates (2017, methodology): “To this end, different techniques have been proposed that score parameters based on their importance (Hutter, Hoos, and Leyton-Brown 2013; 2014; Siegmund et al. 2015; Fawcett and Hoos 2016).”