Sign in

Jascha Sohl-Dickstein

@jascha.sohldickstein.com
7.5K followers 1.4K following 18 posts

Recently a principal scientist at Google DeepMind. Joining Anthropic. Most (in)famous for inventing diffusion models. AI + physics + neuroscience + dynamical systems.

PostsRepliesMedia
Jascha Sohl-Dickstein @jascha.sohldickstein.com · 12/02/2024
Have you ever done a dense grid search over neural network hyperparameters? Like a *really dense* grid search? It looks like this (!!). Blueish colors correspond to hyperparameters for which training converges, redish to those for which training diverges. Even better, a video: vimeo.com/903855670
Examples of fractals resulting from neural network training in a variety of experimental configurations
714031
Jascha Sohl-Dickstein @jascha.sohldickstein.com · 07/11/2023
These Levels of AGI provide a rough framework to quantify the performance, generality, and autonomy of AGI models and their precursors. We hope they help compare models, assess risk, and measure progress along the path to AGI. (AlphaGo is a Level 4 "Virtuouso" Narrow AI)
110
Jascha Sohl-Dickstein @jascha.sohldickstein.com · 07/11/2023
Levels of AGI: Operationalizing Progress on the Path to AGI Levels of Autonomous Driving are extremely useful, for communicating capabilities, setting regulation, and defining goals in self driving. We propose analogous Levels of *AGI*. (ChatGPT is a Level 1 "Emerging" AGI)
170