Discriminative representation loss (DRL) for continual learning: A solution for CLVision challenge

The use of episodic memories in continual learning is an efficient way to prevent the phenomenon of catastrophic forgetting. In recent studies, several gradient-based approaches have been developed to make more efficient use of compact episodic memories, which constrain the gradients resulting from new samples with gradients from memorized samples. In this paper, we propose a method for decreasing the diversity of gradients through an extra optimization objective that we call Discriminative Representation Loss, instead of directly re-projecting the gradients. Our methods show promising performance with relatively cheap computational cost on all the three tracks of the CLVision challenge 1.

[1]  Tom Diethe,et al.  Facilitating Bayesian Continual Learning by Natural Gradients and Stein Gradients , 2019, ArXiv.

[2]  Yu-Hen Hu.  Neural Networks for Signal Processing IX : proceedings of the 1999 IEEE Signal Processing Society Workshop , 1999 .

[3]  David Rolnick,et al.  Experience Replay for Continual Learning , 2018, NeurIPS.

[4]  Gerald Tesauro,et al.  Learning to Learn without Forgetting By Maximizing Transfer and Minimizing Interference , 2018, ICLR.

[5]  Yee Whye Teh,et al.  Progress & Compress: A scalable framework for continual learning , 2018, ICML.

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

[7]  Richard E. Turner,et al.  Variational Continual Learning , 2017, ICLR.

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

[9]  Marc'Aurelio Ranzato,et al.  Efficient Lifelong Learning with A-GEM , 2018, ICLR.

[10]  Yoshua Bengio,et al.  Gradient based sample selection for online continual learning , 2019, NeurIPS.

[11]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[12]  Jiwon Kim,et al.  Continual Learning with Deep Generative Replay , 2017, NIPS.

[13]  Surya Ganguli,et al.  Continual Learning Through Synaptic Intelligence , 2017, ICML.

[14]  Marc'Aurelio Ranzato,et al.  Gradient Episodic Memory for Continual Learning , 2017, NIPS.

[15]  Davide Maltoni,et al.  CORe50: a New Dataset and Benchmark for Continuous Object Recognition , 2017, CoRL.

[16]  Yi-Ming Chan,et al.  Compacting, Picking and Growing for Unforgetting Continual Learning , 2019, NeurIPS.

[17]  Marc'Aurelio Ranzato,et al.  On Tiny Episodic Memories in Continual Learning , 2019 .

[18]  Kilian Q. Weinberger,et al.  Distance Metric Learning for Large Margin Nearest Neighbor Classification , 2005, NIPS.

[19]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[20]  Jinlong Liu,et al.  Understanding Why Neural Networks Generalize Well Through GSNR of Parameters , 2020, ICLR.

[21]  B. Scholkopf,et al.  Fisher discriminant analysis with kernels , 1999, Neural Networks for Signal Processing IX: Proceedings of the 1999 IEEE Signal Processing Society Workshop (Cat. No.98TH8468).