Deep Learning - Loss and Optimization Part 3

This video discusses details on optimization and different options in gradient descent procedure such as momentum and ADAM.

References

[1] Christopher M. Bishop. Pattern Recognition and Machine Learning (Information Science and Statistics). Secaucus, NJ, USA: Springer-Verlag New York, Inc., 2006.
[2] Anna Choromanska, Mikael Henaff, Michael Mathieu, et al. “The Loss Surfaces of Multilayer Networks.” In: AISTATS. 2015.
[3] Yann N Dauphin, Razvan Pascanu, Caglar Gulcehre, et al. “Identifying and attacking the saddle point problem in high-dimensional non-convex optimization”. In: Advances in neural information processing systems. 2014, pp. 2933–2941.
[4] Yichuan Tang. “Deep learning using linear support vector machines”. In: arXiv preprint arXiv:1306.0239 (2013).
[5] Sashank J. Reddi, Satyen Kale, and Sanjiv Kumar. “On the Convergence of Adam and Beyond”. In: International Conference on Learning Representations. 2018.
[6] Katarzyna Janocha and Wojciech Marian Czarnecki. “On Loss Functions for Deep Neural Networks in Classification”. In: arXiv preprint arXiv:1702.05659 (2017).
[7] Jeffrey Dean, Greg Corrado, Rajat Monga, et al. “Large scale distributed deep networks”. In: Advances in neural information processing systems. 2012, pp. 1223–1231.
[8] Maren Mahsereci and Philipp Hennig. “Probabilistic line searches for stochastic optimization”. In: Advances In Neural Information Processing Systems. 2015, pp. 181–189.
[9] Jason Weston, Chris Watkins, et al. “Support vector machines for multi-class pattern recognition.” In: ESANN. Vol. 99. 1999, pp. 219–224.
[10] Chiyuan Zhang, Samy Bengio, Moritz Hardt, et al. “Understanding deep learning requires rethinking generalization”. In: arXiv preprint arXiv:1611.03530 (2016).

Further Reading:
A gentle Introduction to Deep Learning