REVIEW 2 cited by
The centered convex body whose marginals have the heaviest tails
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
The centered convex body whose marginals have the heaviest tails
read the original abstract
Given any real numbers $1<p<q$, we study the norm ratio (i.e. the ratio between the $q$-norm and the $p$-norm) of marginals of centered convex bodies. We first show that some marginal of the simplex maximizes said ratio in the class of $n$-dimensional centered convex bodies. We then pass to the dimension independent (i.e. log-concave) case where we find a 1-parameter family of random variables in which the maximum ratio must be attained, and find the exact maximizer of the ratio when $p=2$ and $q$ is even. In addition, we find another interesting maximization property of marginals of the simplex involving functions with positive third derivatives.
Forward citations
Cited by 2 Pith papers
-
Training Transformers for KV Cache Compressibility
KV compressibility is a property of learned transformer representations that can be improved by training with KV sparsification, leading to better quality-budget tradeoffs in downstream compression for retrieval, QA, ...
-
Training Transformers for KV Cache Compressibility
Training transformers with KV sparsification during continued pretraining produces representations that admit better post-hoc KV cache compression, improving quality under memory budgets for long-context tasks.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.