1

Causal Inference Despite Limited Global Confounding via Mixture Models
The first known algorithm for $k$-MixBND.
Spencer Gordon, Bijan Mazaheri, Yuval Rabani, Leonard Schulman
PDF Cite Project
Causal Inference Despite Limited Global Confounding via Mixture Models
Source Identification for Mixtures of Product Distributions
We develop the “method of synthetic bits” for solving discretem mixtures of product distributions, giving a exponential time complexity improvement (in the number of sources). The algorithm involves a reduction to the $k$-MixIID case.
Spencer Gordon, Bijan Mazaheri, Yuval Rabani, Leonard Schulman
PDF Cite Project
Source Identification for Mixtures of Product Distributions
Synthesizing New Expertise via Collaboration
Nontransitivite properties of the voting systems are also possible in networks of experts and machine learning models.
Bijan Mazaheri, Siddharth Jain, Jehoshua Bruck
PDF Cite Project
Synthesizing New Expertise via Collaboration
Robust Correction of Sampling Bias using Cumulative Distribution Functions
Importance weighting via probability density function estimation is unstable and requires parameter tuning. We demonstrate that reweighting according to cumulative distirbution functions is stable without parameter tuning.
Bijan Mazaheri, Siddharth Jain, Jehoshua Bruck
PDF Cite Project
Robust Correction of Sampling Bias using Cumulative Distribution Functions