Durk Kingma | search
Click here to view Durk Kingma news from 60+ newspapers.
Bookmark or Share- Diederik P. (Durk) KingmaBrief Bio. I do research on scalable methods for machine learning, with a focus on generative models. My contributions include the Variational Autoencoder (VAE), the Adam optimizer, Glow, and Variational Diffusion Models, but please see Scholar for a more complete list.
- Digit Fantasies by a Deep Generative ModelDigit Fantasies by a Deep Generative Model. Instructions: Dream mode: check 'dream' to let the model fantasize digits. Alternatively, you can wiggle the sliders yourselves to wander through z -space and observe the effects in x -space. Randomize!
- Stochastic Backpropagation, Variational Inference, and Semi-Supervised ...We are (very) interested in inferring a posterior distribution p(z|x) e.g.: Full distribution over model parameters. Enables learning parameters in latent-variable models. p(z|x) = p(x,z)/p(x) most often intractable. Need good approximations.
- Speeding up Gradient-Based Inference and Learning in deep/recurrent ...Speeding up Gradient-Based Inference and Learning in deep/recurrent Bayes Nets with Continuous Latent Variables. (by a neural network equivalence) Preprint: http://arxiv.org/abs/1306.0733. Diederik P. (Durk) Kingma Ph.D. Candidate www.dpkingma.com.
- dpkingma.com-5.117866378e-02 -2.035679051e-02 5.918298621e-03 -1.787239267e-02 -1.651095170e-02 -3.292058780e-02 -5.253796868e-02 -2.845141380e-02 -3.034295975e-02 -2.024209944e-02 -5.0078840
More
