[YPNG] YPNG, Friday 10 Nov 2017, classroom at 24 Hillhouse 11:00-1:00

Sekhar Tatikonda sekhar.tatikonda at yale.edu
Mon Nov 6 09:21:41 EST 2017

Hi Folks,

This Friday in the YPNG seminar Cheng Mao will talk about:

Title: Permutation-based models for ranking from pairwise comparisons

Abstract: Pairwise comparison data arises in various domains, including
recommender systems, website ranking and social choice. There has been
a recent surge of interest in studying permutation-based models, such as
the noisy sorting model and the strong stochastic transitivity model, for
ranking from pairwise comparisons. In this talk, I will focus on
estimation for the noisy sorting model given structured partial
The optimal rate of estimation is established up to a constant (resp. a
logarithmic factor) for pairwise comparisons sampled according to an
Erdős–Rényi graph (resp. certain classes of fixed graphs). In addition,
I will discuss efficient algorithms that generalize the classical Borda
count estimator. The talk is based on the following joint works with
Jonathan Weed, Philippe Rigollet, Ashwin Pananjady, Vidya
Muthukumar, Martin J. Wainwright, Thomas A. Courtade:


Bio: Cheng Mao is a fifth-year graduate student in the Mathematics
Department at MIT, working with Professor Philippe Rigollet. His broad
research interests include statistics theory, machine learning and
Recently, he has been working on matrix and graph estimation problems with
latent permutations. Cheng received his B.S. and M.A degrees in
Mathematics from UCLA.

See you Friday at 11 in the S&DS classroom.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.yale.edu/pipermail/ypng/attachments/20171106/18879ed2/attachment.html 

More information about the YPNG mailing list