Personal Info

Resume

Technical Presentations

A Distributed Gibbs Sampler for LDA using Spark and GraphX at MLCONF 2015, Seattle

• Video
• Presentation

Teaching Probabilistic Graphical Models

I lead a reading group on Advanced Probabilistic Modeling internally in Netflix. Some of my notes and implementations covering Mixture Modeling, Expectation-Maximization, Gibbs Sampling and Variational Inference are linked below:

Notes

• Maximum Likelihood Estimation
• Expectation-Maximization
• Bayesian Mixtures
• MAP estimates
• Posterior Sampling using Gibbs Sampling
• Posterior Sampling using Collapsed Gibbs Sampling
• Computing Posterior Distribution using Variational Inference - 1
• Computing Posterior Distribution using Variational Inference - 2
• Scaling Variational Inference using Stochastic Optimization

IPython Notebooks

• Multinomial Mixture Model inference using EM, Collapsed Gibbs Sampling and Stochastic Variational Inference

Select Publications

• Learning Representations of Hierarchical Slates in Collaborative Filtering, RecSys 2020
• Variational Low Rank Multinomial Models for Collaborative Filtering with Side-information, RecSys 2019
• A Nonparametric Latent Factor Model For Location-Aware Video Recommendations, NIPS Workshop on Practical Bayesian Nonparametrics, 2016