Ghazal Fazelnia
Research Scientist at Spotify Research
Research Interests: Deep Generative Models, Bayesian Inference, Reinforcement Learning, Recommender Systems, Optimization
About
I am a Research Scientist at Spotify Research. My work is focused on developing machine learning models for understanding latent patterns in data using deep generative models, approximate inference, and reinforcement learning. I received my PhD in Electrical Engineering Department and Data Science Institute at Columbia University in 2019 with my thesis "Optimization for Probabilistic Machine Learning" advised by Professor John Paisley. I graduated with my B.Sc. in Electrical Engineering from Sharif University of Technology, Iran, 2013.
Email: firstname at ee dot columbia dot edu
Publication
Zhenwen Dai, Praveen Chandar, Ghazal Fazelnia, Ben Carterette, Mounia Lalmas, "Model Selection for Production System via Automated Online Experiments", NeurIPS, 2020.
Greg Benton, Ghazal Fazelnia, Alice Wang, Ben Carterette, "Trajectory Based Podcast Recommendation", PodRec Workshop, ACM Conference on Recommender Systems (RecSys), 2020.
Ghazal Fazelnia, Mark Ibrahim, Ceena Modarres, Kevin Wu, John Paisley, "Mixed Membership Recurrent Neural Networks for modeling customer purchases", ACM International Conference on AI in Finance, 2020.
Ghazal Fazelnia, John Paisley, "CRVI: Convex Relaxation for Variational Inference", ICML, 2018.
Ghazal Fazelnia, Ramtin Madani, Abdulrahman Kalbat, Javad Lavaei, "Convex Relaxation for Optimal Distributed Control Problem," IEEE Transactions on Automatic Control, 2017.
Salar Fattahi, Ghazal Fazelnia, Javad Lavaei, Murat Arcak, "Transformation of Optimal Centralized Controllers into Near-Globally Optimal Static Distributed Controllers," IEEE Transaction on Automatic Control, 2017.
Ramtin Madani, Somayeh Sojoudi, Ghazal Fazelnia, Javad Lavaei, "Finding Low-rank Solutions of Sparse Linear Matrix Inequalities using Convex Optimization," SIAM Journal on Optimization, 2017.
Salar Fatahi, Ghazal Fazelnia, Javad Lavaei, "Transformation of Optimal Centralized Controllers Into Near-Global Static Distributed Controllers," Proc. of the 54th IEEE Conference on Decision and Control, 2015.
Ghazal Fazelnia, Ramtin Madani, Javad Lavaei, "Convex Relaxation for Optimal Distributed Control Problem," Proc. of the 53rd IEEE Conference on Decision and Control, 2014.
Ramtin Madani, Ghazal Fazelnia, Somayeh Sojoudi, Javad Lavaei, "Low-Rank Solutions of Matrix Inequalities With Applications to Polynomial Optimization and Matrix Completion Problems," Proc. of the 53rd IEEE Conference on Decision and Control, 2014. Best Paper Finalist.
Somayeh Sojoudi, Ramtin Madani, Ghazal Fazelnia, Javad Lavaei, "Graph-Theoretic Algorithms for Polynomial Optimization Problems," Proc. of the 53rd IEEE Conference on Decision and Control, 2014 (Invited tutorial paper).
Abdulrahman Kalbat, Ramtin Madani, Ghazal Fazelnia, Javad Lavaei, "Efficient Convex Relaxation for Stochastic Optimal Distributed Control Problem," Allerton Conference, 2014.