Lunch In Theory This Thursday (12:00, 11/13, GCS 502c)
2 views
Skip to first unread message
Devansh Gupta
Nov 12, 2025, 10:21:05 PM (14 days ago) Nov 12
to usc-t...@googlegroups.com, USC Theory Group, CS Theory Group, Alimohammadi, Yeganeh
Hi all,
Please join us for Lunch in Theory this Thursday, 10/30 at 12:00 in GCS 502c.
Reminder: Please bring your own lunch, as lunch will not be provided.
This week we have a speaker from Marshall, Yeganeh Alimohammadi, giving a talk on ranking models. Please find the title and abstract attached.
We look forward to having you all in the talk.
Best,
Devansh
Title: How to Measure Differences in Rankings: A Data-Driven Extension of the Mallows Model
Abstract: Ranking problems appear across domains, from consumer preferences and product recommendations to sports performance and hiring decisions. Probabilistic ranking models such as the Mallows model provide a principled way to capture uncertainty and infer a consensus order, assuming that observed rankings are noisy versions of an underlying truth. A key modeling challenge, however, lies in choosing the distance function that measures how “far apart” two rankings are, since different distances imply different ways of penalizing ranking swaps.
In this talk, I introduce a generalized framework based on Mallows model that learns this distance metric directly from data. I will show that the model forms an exponential-family distribution on permutations, and that its parameters (the central ranking, dispersion, and learned distance) can be estimated consistently via maximum likelihood with asymptotic normality guarantees. On the algorithmic side, I present a polynomial-time approximation scheme (PTAS) for efficient sampling and partition-function estimation. Finally, I will discuss empirical validation on real datasets, demonstrating how learning the distance metric leads to more accurate predictions and interpretable insights about ranking behavior.
Abstract: Ranking problems appear across domains, from consumer preferences and product recommendations to sports performance and hiring decisions. Probabilistic ranking models such as the Mallows model provide a principled way to capture uncertainty and infer a consensus order, assuming that observed rankings are noisy versions of an underlying truth. A key modeling challenge, however, lies in choosing the distance function that measures how “far apart” two rankings are, since different distances imply different ways of penalizing ranking swaps.
In this talk, I introduce a generalized framework based on Mallows model that learns this distance metric directly from data. I will show that the model forms an exponential-family distribution on permutations, and that its parameters (the central ranking, dispersion, and learned distance) can be estimated consistently via maximum likelihood with asymptotic normality guarantees. On the algorithmic side, I present a polynomial-time approximation scheme (PTAS) for efficient sampling and partition-function estimation. Finally, I will discuss empirical validation on real datasets, demonstrating how learning the distance metric leads to more accurate predictions and interpretable insights about ranking behavior.
Devansh Gupta
Nov 13, 2025, 2:37:19 PM (13 days ago) Nov 13
to usc-t...@googlegroups.com, USC Theory Group, CS Theory Group, Alimohammadi, Yeganeh
Hi everyone,
Please find the meet link for the talk attached: https://usc.zoom.us/j/91362198629?pwd=npZ13f5Q2v74Pmoe6tyXwjvJ0V9s5b.1.
Best,
Devansh
Reply all
Reply to author
Forward
0 new messages