Machine Learning for Physicists

ver. 20171218

■ Lecturer I: Prelude   조정효 Junghyo Jo (KIAS)

1. Perceptron

2. Recurrent neural network

3. Boltzmann machine

4. Causality inference

■ Lecture II: Machine Learning 101  노영균 Yung-Kyun Noh (SNU)

1. Probability theory

2. Bayes classifier and regression

3. Directed and undirected graphical models

4. Inference using Kalman filter

5. Latent variable models

■ Lecturer III: Machine Learning in Practice  안강헌 Kang-Hun Ahn (CNU)

1. Convolutional neural network

2. Generative adversarial network

3. Python and TensorFlow^TM

■ Special Lecture: Information Dynamics   Chaoming Song (Miami)

1. Least action principle vs. maximum log-likelihood

2. Stochastic process, a mini primer

3. Reinforced Poisson process and citation dynamics

4. Substitutional systems

* General References

1. David Rumelhart, Geoffrey Hinton and Ronald Williams (1986),
   Learning representations by back-propagating errors, Nature 323: 533-536.
   
   
2. Shun-ichi Amari, Koji Kurata and Hiroshi Nagaoka (1992), Information geometry
   of Boltzmann machines, IEEE Transactions on Neural Networks 3: 260-271.
   
   
3. Herbert Jaeger (2002), A tutorial on training recurrent neural networks, covering
   BPTT, RTRL, EKF and the “echo state network” approach, GMD Report 159,
   Fraunhofer Institute AIS.
   
   
4. Geoffrey Hinton (2012), A Practical Guide to Training Restricted Boltzmann Machines.
   In: Montavon G., Orr G.B., Muller KR. (eds) Neural Networks: Tricks of the Trade.
   Lecture Notes in Computer Science, vol 7700. Springer, Berlin, Heidelberg.
   
   
5. Geoffrey Hinton and Ruslan Salakhutdinov (2006), Reducing the dimensionality of
   data with neural networks, Science 313: 504-507.
   
   
6. Ilya Sutskever, Geoffrey Hinton and Graham Taylor (2009), The Recurrent Temporal
   Restricted Boltzmann Machine. Advances in Neural Information Processing
   Systems 21, MIT Press, Cambridge, MA.