Adaptively Growing Hierarchical Mixtures of Experts

How researchers see this paper

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

A modeling framework for adaptive lifelong learning with transfer and savings through gating in the prefrontal cortex (2020, background): “1a) widely found in machine learning applications [19],[20],[21],[22].”
Inference for Subgroup Analysis With a Structured Logistic-Normal Mixture Model (2015, methodology): “Bayesian methods for selecting the number of experts were suggested by Peng, Jacobs, and Tanner (1996) and Ueda and Ghahramani (2002), whereas Fritsch, Finke, and Waibel (1997) considered a grow-and-prune strategy of the model for the same purpose.”
On Anytime Learning at Macroscale (2021, methodology): “Towards this goal, in this work we consider a simple extension of the one layer MoE by [Fritsch et al., 1996].”
Twenty Years of Mixture of Experts (2012, background): “To this end, the studies that aim to arrive at the optimal structure of the tree can be divided into four main categories as follows: growing models [2], [65], [66]; pruning models [49], [67], [68]; exhaustive search [45]; and Bayesian models [10], [18], [54], [69].”
Fully-Automatic Bayesian Piecewise Sparse Linear Models (2014, methodology): “[16, 17] proposed HME learning algorithms using growing strategies rather than EM-iteration.”