A quantitative measure of mode-seeking vs. mode-covering preferences
Another view
Another view
Stein discrepancy is more mode-seeking than reverse KL
An impossibility result
What does mode-seeking vs. mode-covering behavior in distributional divergences really mean?
A brainteaser
When AI agents outnumber voters, how should we elect an AI government?
How do modern deep generative models produce high-quality outputs?
Is data sacred?
A speculative argument
A brief mathematical result
Exploring connections between diffusion models, alphafold’s recycling, structured prediction energy networks, and manifold learning.
An introduction
An introduction.
KL behaves differently in discrete spaces than in continuous spaces.