4 papers
Differentially Private Language Generation and Identification in the Limit
Anay Mehrotra, Grigoris Velegkas, Xifan Yu +1
We initiate the study of language generation in the limit, a model recently introduced by Kleinberg and Mullainathan [KM24], under the constraint of differential privacy. We consid…
Stable Algorithms Lower Bounds for Estimation
Xifan Yu, Ilias Zadik
In this work, we show that for all statistical estimation problems, a natural MMSE instability (discontinuity) condition implies the failure of stable algorithms, serving as a vers…
Language Generation with Infinite Contamination
Anay Mehrotra, Grigoris Velegkas, Xifan Yu +1
We study language generation in the limit, where an algorithm observes an adversarial enumeration of strings from an unknown target language and must eventually generate new, u…
Statistical inference of a ranked community in a directed graph
Dmitriy Kunisky, Daniel A. Spielman, Alexander S. Wein +1
We study the problem of detecting or recovering a planted ranked subgraph from a directed graph, an analog for directed graphs of the well-studied planted dense subgraph model. We…