Statistical guarantees for the EM algorithm: From population to sample-based analysis · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/1408.2156/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTStatistical guarantees for the EM algorithm: From population to sample-based analysis — mentioned by mturmonhttps://news.ycombinator.com/item?id=16223492urn:hackerpapers:mention:1408.2156:16223492Wed, 24 Jan 2018 15:41:23 GMT<p>There has been a big question about the place of convex relaxations for solving learning problems, and this responds to that question. That is, you have a learning problem with some fit criterion that is non-convex. Maybe you choose a different objective function, or different parameterization of the problem, so that it is convex. Then you can use a convex solver to find the guaranteed optimum of the &quot;wrong&quot; problem. Using convex relaxations has been a trend for perhaps 20 years - it avoids a whole bunch of nagging &quot;what if my solution is just a local max&quot; questions. Or, you can use older…</p><p><a href="https://hackerpaper.kaipfister.de/p/1408.2156/">Paper and discussion history</a>. A mention is not an endorsement.</p>math.ST