This work proposes a novel approach that mitigates noise in task-relevant subspaces of the data representation by identifying task-critical subspaces via the Jacobian of a public downstream model, selectively attenuates noise along these directions, and reshapes the isotropic noise of standard LDP mechanisms into an an...
Young-Mok Ha, Viktor Schlegel, Yi-Dan Sun et al.· arXiv.org· 0 citations
The audit shows that synthetic releases leak subgroup membership and that prior attacks systematically underestimate this leakage, and that DP is effective at the aggregate level: it substantially reduces average leakage at every budget the authors test.
Yi-Dan Sun, Viktor Schlegel, Srinivasan Nandakumar et al.· 0 citations
This work introduces HealthLoopQA, a comprehensive diagnostic benchmark for evaluating LLM reasoning over continuous diabetes monitoring data, and identifies a broader phenomenon of In-context Laziness under long-context prompting, highlighting critical open challenges in deploying LLMs for rigorous long-horizon medica...
Yu-Chen Niu, Ya-Nan Ma, Srinivasan Nandakumar et al.· 0 citations
Compared to existing DP text generation baselines, Term2Note substantially improves both fidelity and utility, without relying on label distribution assumptions, highlighting its effectiveness as a practical privacy-preserving alternative to real clinical notes.
Yuping Wu, Viktor Schlegel, Warren Del-Pinto et al.· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
