"Stabilizing Generative Adversarial Network Training: A Survey" -- imho, training GANs can be extremely frustrating; this is a nice paper to keep handy for these occasions: https://t.co/OsbnxXVhb4
— Sebastian Raschka (@rasbt) October 3, 2019
"Stabilizing Generative Adversarial Network Training: A Survey" -- imho, training GANs can be extremely frustrating; this is a nice paper to keep handy for these occasions: https://t.co/OsbnxXVhb4
— Sebastian Raschka (@rasbt) October 3, 2019
Do you formally know Monte-Carlo and TD learning, but don't intuitively understand the difference? This is for you.https://t.co/2DR75rK40u (with @samgreydanus) pic.twitter.com/6RwsBjFbU9
— Chris Olah (@ch402) October 1, 2019
Learn how we built our distributed k-means #machinelearning algorithm. Read how #GPUs + multi-node RAPIDS can deliver a 200x performance improvement over CPUs, while keeping a familiar API. https://t.co/GtSijhliOS
— RAPIDS AI (@rapidsai) September 26, 2019
What a great article. I don't think I've come across too much on the subject of rigorous software testing for ml models. And technical articles that are as well written and clear as this are rare. https://t.co/qSH9AMkXPo
— Jason Antic (@citnaj) September 24, 2019
I came across this really awesome explanation and comparison of a variety of methods to estimate predictive intervals from neural networks (in PyTorch). A great starting point if you’re thinking about how to add uncertainty to your model. https://t.co/cpdUfP3dXh pic.twitter.com/fzs8obZST4
— Sean J. Taylor (@seanjtaylor) September 19, 2019
We get a LOT of questions about how best to use Dask efficiently.
— Dask (@dask_dev) September 17, 2019
We now curate lists of Best Practices here:https://t.co/gWHVo0qisL
There is one page for the entire project, and one page for each of Arrays, DataFrames, and Delayed.
This is the best distillation of recent (and old!) research on batchnorm I've seen.
— Jeremy Howard (@jeremyphoward) September 11, 2019
There is so much to learn about training mechanics by studying this thread and the links it contains. https://t.co/a1PeCy7M1s
So we have given precise experimental meaning to the statement that 'internal covariate shift' limits LRs and that BN works by preventing this...
— David Page (@dcpage3) September 11, 2019
...matching the intuition of the original paper!
Recent papers have studied the Hessian of the loss for deep nets experimentally:
— David Page (@dcpage3) September 11, 2019
(@leventsagun et al) https://t.co/JNJKeqZyvZ, https://t.co/Wbk3sSbIbr
(Papyan) https://t.co/l4QcB85nir.
(@_ghorbani et al) https://t.co/VUxknF5QkM compare what happens with and without BN.
Let's get started.
— David Page (@dcpage3) September 11, 2019
What does the original paper have to say? pic.twitter.com/SUehxk58Jt
Over the years, I've watched conditional renormalization grow from a style transfer hack to a key mechanism in recent #MachineLearning results; in this post I chart out how the idea evolved from humble beginnings into a flexible and important techniquehttps://t.co/OXCVnpaLnP
— Cody Wild (@decodyng) September 8, 2019
Christ, @ryxcommar took the incredible thread following my gripe on sklearn's surprising defaults for LogisticRegression (L2 reg, C=1.0) and turned it into a marvelous blog post, fleshing out why precisely this trend (in many ML libraries) is dangerous.
— Zachary Lipton (@zacharylipton) September 1, 2019
→https://t.co/CNbxWfoe6N https://t.co/IAmuBZ9Uov