Animation: Variational Autoencoder
A variational autoencoder is a type of neural network that learns to compress (encode) data, in such a way that one can later randomly sample data.
Here we show training of a variational autoencoder on a data set of wave packets that have been placed with a random location and amplitude. Upper left: test set, with VAE decoder results shown in orange; upper right: latent space distribution, with color coding for actual location (left) or amplitude (right) of any data point; lower right: latent space distributions for the few samples from the test set (in each case, the Gaussian spread in latent space is also indicated); lower left: loss evolution.
2021 by Florian Marquardt. For the full explanation, watch https://www.youtube.com/watch?v=bSta4s439-I .
This animation is part of the online lecture series "Advanced Machine Learning for Physics, Science, and Artificial Scientific Discovery". See the website https://pad.gwdg.de/s/2021_AdvancedMachineLearningForScience# and the channel with the full lecture videos: https://www.youtube.com/playlist?list=PLemsnf33Vij4-kv-JTjDthaGUYUnQbbws.
Видео Animation: Variational Autoencoder канала Florian Marquardt
Here we show training of a variational autoencoder on a data set of wave packets that have been placed with a random location and amplitude. Upper left: test set, with VAE decoder results shown in orange; upper right: latent space distribution, with color coding for actual location (left) or amplitude (right) of any data point; lower right: latent space distributions for the few samples from the test set (in each case, the Gaussian spread in latent space is also indicated); lower left: loss evolution.
2021 by Florian Marquardt. For the full explanation, watch https://www.youtube.com/watch?v=bSta4s439-I .
This animation is part of the online lecture series "Advanced Machine Learning for Physics, Science, and Artificial Scientific Discovery". See the website https://pad.gwdg.de/s/2021_AdvancedMachineLearningForScience# and the channel with the full lecture videos: https://www.youtube.com/playlist?list=PLemsnf33Vij4-kv-JTjDthaGUYUnQbbws.
Видео Animation: Variational Autoencoder канала Florian Marquardt
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Lecture 4: Loss functions. Overfitting. Dropout. Adaptive Gradient Descent. Convolutional networks.
Lecture 26: Active Learning for Network Training: Uncertainty Sampling and other approaches.
Lecture 23: Reinforcement Learning - Policy Gradient and Q-Learning.
Lecture 14: Boltzmann Machines (General Theory).
Lecture 19: Graph Neural Networks. Attention Mechanisms (Basics).
Lecture 10: Inductive Bias. Fisher Information. Information Geometry.
Moderne Physik: "Auf der Jagd nach kosmischen Teilchen." (Prof. Anna Nelles)
Lecture 21: Transformers (and examples). Implicit Layers.
Lecture 12: Mutual Information. Learning Probability Distributions. Normalizing Flows.
Talk: Discovering feedback strategies for open quantum systems via deep reinforcement learning
Machine Learning for Physicists (Lecture 3): Training networks, Keras, Image recognition
Lecture 16: Variational Autoencoder. Generative Adversarial Networks.
Lecture 11: Natural Gradient. Kullback-Leibler Divergence. Mutual Information.
Lecture 15: Restricted Boltzmann Machines. Conditional Sampling. Variational Autoencoder.
Machine Learning for Physicists (Lecture 5): Principal Component Analysis, t-SNE, Adam etc., ...
Lecture 25: Reinforcement Learning: Continuous actions. Model-based. Monte Carlo Tree Search.
Lecture 7: Contractive Autoencoder. Shannon's Information Theory: Compression and Information.
Lecture 27: Bayesian Optimal Experimental Design. Active Learning: Gaussian Processes and Networks.
Lecture 22: Implicit Layers. Hamiltonian and Lagrangian Networks. Reinforcement Learning Overview.
Lecture 20: Attention. Differentiable Neural Computer. Transformers.