2020

When Ensembling Smaller Models is More Efficient than Single Large Models

Kondratyuk, Dan, Tan, Mingxing, Brown, Matthew et al.

Understand

Ensembling is a simple and popular technique for boosting evaluation performance by training multiple models (e.g., with different initializations) and aggregating their predictions.

  • This approach is commonly reserved for the largest models, as it is commonly held that increasing the model size provides a more substantial reduction in error than ensembling smaller models.
  • However, we show results from experiments on CIFAR-10 and ImageNet that ensembles can outperform single models with both higher accuracy and requiring fewer total FLOPs to compute, even when those individual models' weights and hyperparameters are highly optimized.
  • Furthermore, this gap in improvement widens as models become large.

Built on

  • Neural network ensembles

    Lars Kai Hansen and Peter Salamon · 1990

    Earlier work this paper cites.

  • When networks disagree: Ensemble methods for hybrid neural networks

    Michael P Perrone and Leon N Cooper · 1992

    Earlier work this paper cites.

  • Neural network ensembles, cross validation, and active learning

    Anders Krogh and Jesper Vedelsby · 1995

    Earlier work this paper cites.

  • Bagging predictors

    Leo Breiman · 1996

    Earlier work this paper cites.

  • Engineering multiversion neural-net systems

    Derek Partridge and William B Yates · 1996

    Earlier work this paper cites.

Similar

  • Learning multiple layers of features from tiny images

    Alex Krizhevsky, Geoffrey Hinton, et al · 2009

    Cited alongside, same era.

  • Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton · 2012

    Cited alongside, same era.

  • Distilling the knowledge in a neural network

    Original

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean · 2015

    Cited alongside, same era.

  • Imagenet large scale visual recognition challenge

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al · 2015

    Cited alongside, same era.

Then

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…