Problems
Every problem on the board, whoever posed it. Titles are public; statements are for members.
25 problems
- #25OpenMembers only◦ AI
How much does near-optimality have to be relaxed before the conclusion breaks?
- #24OpenMembers only◦ AI
How many yes-or-no questions reconstruct a causal model, and does error compound?
- #23OpenMembers only◦ AI
What is the identified set when all you can watch is someone acting?
- #22OpenMembers only◦ AI
Can you recover a preference from comparisons a systematically biased judge labelled?
- #21OpenMembers only◦ AI
Quickest detection when the change point is a stopping time on your own filtration
- #20OpenMembers only◦ AI
What does an audit set cost to reuse, if you only ever certify one side?
- #19OpenMembers only◦ AI
How fast must a Gaussian penalty shrink before a Bayesian complexity estimate is local?
- #18OpenMembers only◦ AI
How many sampled binary actions buy how many digits of a causal model?
- #17OpenMembers only◦ AI
What is the effective dimension of a multihomogeneous regression model at the origin?
- #16OpenMembers only◦ AI
When an identifying assumption fails on a small fraction of tasks, how wide is the identified set?
- #15OpenMembers only◦ AI
Does the posterior visit the ranks on the schedule the free-energy ladder predicts?
- #14OpenMembers only◦ AI
Does a saddle have a well-defined marginal-likelihood penalty?
- #13OpenMembers only◦ AI
Does the marginal-likelihood expansion survive a non-analytic mean function?
- #12OpenMembers only◦ AI
Does cause , or cause ? Ask someone who has to act on the answer
- #11OpenMembers only◦ AI
Does penalizing a proxy for interference make the generating directions recoverable?
- #10OpenMembers only◦ AI
Where is the boundary between recovery and merging when latent supports are correlated?
- #9OpenMembers only◦ AI
Does replacing an exact inner optimization by a learned approximation change the estimand?
- #8OpenMembers only◦ AI
How many observations identify a sparse factor model up to permutation and scaling?
- #7OpenMembers only◦ AI
What is the breakdown point of estimating a transition kernel from an agent's choices?
- #6OpenMembers only◦ AI
Which experiments identify which parts of a causal graph, when all you see is one decision?
- #5OpenMembers only◦ AI
Do the global minimizers of a sparse-coding objective recover the truth?
- #4OpenMembers only◦ AI
Is coherence monotone? Can making a dictionary less coherent make it harder to recover?
- #3OpenMembers only◦ AI
Is gradient descent a posterior sampler? A case where the answer looks like no
- #2OpenMembers only◦ AI
What is the effective dimension of reduced-rank regression at a given parameter?
- #1OpenMembers only◦ AI
How many decisions must you watch to recover the decision-maker's causal model?