An Empirical Study of Example Forgetting during Deep Neural Network Learning

Inspired by the phenomenon of catastrophic forgetting, we investigate the learning dynamics of neural networks as they train on single classification tasks. Our goal is to understand whether a related phenomenon occurs when data does not undergo a clear distributional shift. We define a “forgetting event” to have occurred when an individual training example transitions from being classified correctly to incorrectly over the course of learning. Across several benchmark data sets, we find that: (i) certain examples are forgotten with high frequency, and some not at all; (ii) a data set’s (un)forgettable examples generalize across neural architectures; and (iii) based on forgetting dynamics, a significant fraction of examples can be omitted from the training data set while still maintaining state-of-the-art generalization performance.

[1]  Hugo Larochelle,et al.  Optimization as a Model for Few-Shot Learning , 2016, ICLR.

[2]  Andrew M. Saxe,et al.  High-dimensional dynamics of generalization error in neural networks , 2017, Neural Networks.

[3]  Razvan Pascanu,et al.  Overcoming catastrophic forgetting in neural networks , 2016, Proceedings of the National Academy of Sciences.

[4]  Huan Wang,et al.  Identifying Generalization Properties in Neural Networks , 2018, ArXiv.

[5]  Andrew McCallum,et al.  Active Bias: Training More Accurate Neural Networks by Emphasizing High Variance Samples , 2017, NIPS.

[6]  Yoshua Bengio,et al.  A Closer Look at Memorization in Deep Networks , 2017, ICML.

[7]  Michael McCloskey,et al.  Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem , 1989 .

[8]  Chico Q. Camargo,et al.  Deep learning generalizes because the parameter-function map is biased towards simple functions , 2018, ICLR.

[9]  François Fleuret,et al.  Not All Samples Are Created Equal: Deep Learning with Importance Sampling , 2018, ICML.

[10]  Daphne Koller,et al.  Self-Paced Learning for Latent Variable Models , 2010, NIPS.

[11]  Nathan Srebro,et al.  The Implicit Bias of Gradient Descent on Separable Data , 2017, J. Mach. Learn. Res..

[12]  Carla E. Brodley,et al.  Identifying Mislabeled Training Data , 1999, J. Artif. Intell. Res..

[13]  Yong Jae Lee,et al.  Learning the easy things first: Self-paced visual category discovery , 2011, CVPR 2011.

[14]  Percy Liang,et al.  Understanding Black-box Predictions via Influence Functions , 2017, ICML.

[15]  Yoshua Bengio,et al.  Scaling learning algorithms towards AI , 2007 .

[16]  Ryota Tomioka,et al.  In Search of the Real Inductive Bias: On the Role of Implicit Regularization in Deep Learning , 2014, ICLR.

[17]  Graham W. Taylor,et al.  Improved Regularization of Convolutional Neural Networks with Cutout , 2017, ArXiv.

[18]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[19]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[20]  Jason Weston,et al.  Curriculum learning , 2009, ICML '09.

[21]  Yi Zhou,et al.  Convergence of SGD in Learning ReLU Models with Separable Data , 2018, ArXiv.

[22]  Jason Yosinski,et al.  Measuring the Intrinsic Dimension of Objective Landscapes , 2018, ICLR.

[23]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[24]  Nikos Komodakis,et al.  Wide Residual Networks , 2016, BMVC.

[25]  Yuanzhi Li,et al.  An Alternative View: When Does SGD Escape Local Minima? , 2018, ICML.

[26]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[27]  George H. John.  Robust Decision Trees: Removing Outliers from Databases , 1995, KDD.

[28]  Tong Zhang,et al.  Stochastic Optimization with Importance Sampling for Regularized Loss Minimization , 2014, ICML.

[29]  Samy Bengio,et al.  Understanding deep learning requires rethinking generalization , 2016, ICLR.

[30]  David Barber,et al.  Online Structured Laplace Approximations For Overcoming Catastrophic Forgetting , 2018, NeurIPS.

[31]  Yann LeCun,et al.  The mnist database of handwritten digits , 2005 .

[32]  Tom Schaul,et al.  Prioritized Experience Replay , 2015, ICLR.

[33]  Jorge Nocedal,et al.  On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima , 2016, ICLR.

[34]  Joan Bruna,et al.  Training Convolutional Networks with Noisy Labels , 2014, ICLR 2014.

[35]  Tao Qin,et al.  Learning What Data to Learn , 2017, ArXiv.

[36]  Li Fei-Fei,et al.  MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels , 2017, ICML.

[37]  Jürgen Schmidhuber,et al.  Flat Minima , 1997, Neural Computation.

[38]  Jonghyun Choi,et al.  ScreenerNet: Learning Curriculum for Neural Networks , 2018, ArXiv.

[39]  Yoshua Bengio,et al.  On the Learning Dynamics of Deep Neural Networks , 2018, ArXiv.