Timezone: »

 
Poster
Learning Structured Distributions From Untrusted Batches: Faster and Simpler
Sitan Chen · Jerry Li · Ankur Moitra

Tue Dec 08 09:00 AM -- 11:00 AM (PST) @ Poster Session 1 #517
in Poster Session 2 »

We revisit the problem of learning from untrusted batches introduced by Qiao and Valiant [QV17]. Recently, Jain and Orlitsky [JO19] gave a simple semidefinite programming approach based on the cut-norm that achieves essentially information-theoretically optimal error in polynomial time. Concurrently, Chen et al. [CLM19] considered a variant of the problem where μ is assumed to be structured, e.g. log-concave, monotone hazard rate, t-modal, etc. In this case, it is possible to achieve the same error with sample complexity sublinear in n, and they exhibited a quasi-polynomial time algorithm for doing so using Haar wavelets.

In this paper, we find an appealing way to synthesize the techniques of [JO19] and [CLM19] to give the best of both worlds: an algorithm which runs in polynomial time and can exploit structure in the underlying distribution to achieve sublinear sample complexity. Along the way, we simplify the approach of [JO19] by avoiding the need for SDP rounding and giving a more direct interpretation of it through the lens of soft filtering, a powerful recent technique in high-dimensional robust estimation. We validate the usefulness of our algorithms in preliminary experimental evaluations.

Author Information

Sitan Chen (MIT)
Jerry Li (Microsoft)
Ankur Moitra (MIT)

More from the Same Authors