Nikhil Rao
Senior ML Scientist
Amazon
My research interests lie in developing models and efficient algorithms for learning from high dimensional structured data, and their applications in information retrieval, natural language processing and detecting statistically rare events. I'm currently applying these principles at Amazon Search, where I build large scale ML algorithms on graphs. I was previously a Machine Learning Researcher in the Technicolor AI Lab, and prior to that, a recipient of the ICES Postdoctoral Fellowship at the University of Texas at Austin. I received my PhD in Electrical and Computer Engineering from the University of Wisconsin-Madison.
Publications:
ANTHEM: Attentive Hyperbolic Entity Model for Product Search (WSDM 2022), with Nurendra Choudhary, Sumeet Katariya, Karthik Subbian and Chandan Reddy.
Probabilistic Entity Representation Model for Chain Reasoning over Knowledge Graphs (NeurIPS 2021), with Nurendra Choudhary, Sumeet Katariya, Karthik Subbian and Chandan Reddy. (Code).
Bipartite Dynamic Representations for Abuse Detection (KDD 2021), with Andrew Wang, Rex Ying, Pan Li, Karthik Subbian and Jure Leskovec (Code coming soon)
Self Supervised Hyperboloid Representations from Logical Queries over Knowledge Graphs (WWW 2021), with Nurendra Choudhary, Sumeet Katariya, Karthik Subbian and Chandan Reddy. (Code).
Finding Needles in Heterogenous Haystacks (IAAI 2021), with Bijaya Adhikari, Liangyue Li and Karthik Subbian
Regularized Graph Convolutional Networks for Short Text Classification (COLING 2020), with Kshitij Tayal, Saurabh Agrawal, Xiaowei Jia, Karthik Subbian and Vipin Kumar
Learning Robust Models for Product Search (ACL 2020), with Thanh Nguyen and Karthik Subbian
Graph DNA: Deep Neighborhood Aware Graph Encoding for Collaborative Filtering (AISTATS 2020), with Liwei Wu, Hsiang-Fu Yu, James Sharpnack and Cho-Jui Hsieh
Scalable Feature Selection for (Multitask) Gradient Boosting Machines (AISTATS 2020), with Cuize Han, Daria Sorokina and Karthik Subbian. (Blog Post)
Language Agnostic Representation Learning for Product Search (WSDM 2020), with Aman Ahuja, Sumeet Katariya, Karthik Subbian and Chandan Reddy. (Blog Post, Press)
Online Bayesian Learning for E-Commerce Query Reformulation (NeuRIPS Bayesian Deep Learning workshop 2019), with Gaurush Hiranandani, Sumeet Katariya and Karthik Subbian
Short Text Classification using Graph Convolutional Networks (NeuRIPS Graph Workshop 2019), with Kshitij Tayal, Saurabh Agrawal and Karthik Subbian
Identifying Facet Mismatches In Search Via Micrographs (CIKM 2019), with Sriram Srinivasan, Karthik Subbian and Lise Getoor.
A Sparse Topic Model for Extracting Aspect-Specific Summaries from Online Reviews (WWW 2018), with Vineeth Rakesh, Weicong Ding, Aman Ahuja, Yifan Sun and Chandan Reddy [CODE]
Discovery of Evolving Semantics through Dynamic Word Embedding Learning (WSDM 2018), with Zijun Yao, Weicong Ding, Yifan Sun and Hui Xiong [CODE]
Matrix Factorization with Side and Higher Order Information (KDD MLG workshop 2017), with Vatsal Shah and Weicong Ding
The Group k-Support Norm for Learning with Structured Sparsity (ICASSP 2017), with Miroslav Dudik and Zaid Harchaoui
On Learning High Dimensional Structured Single Index Models (AAAI 2017), with Ravi Ganti, Laura Balzano, Rebecca Willett and Robert Nowak
Classification with the Sparse Group Lasso (IEEE Trans. Signal Processing 2016), with Robert Nowak, Chris Cox and Tim Rogers [CODE]
High Dimensional Time Series Prediction with Missing Values (NIPS 2016) with Hsiang Fu Yu and Inderjit Dhillon
Structured Sparse Regression via Greedy Hard Thresholding (NIPS 2016) with Prateek Jain and Inderjit Dhillon
Goal-Directed Inductive Matrix Completion (KDD 2016) with Si Si, Kai-Yang Chiang, Cho-Jui Hsieh and Inderjit Dhillon
Forward - Backward Greedy Algorithms for Atomic Norm Regularization (IEEE Trans. Signal Processing, 2015) with Parikshit Shah and Stephen Wright [CODE]
Sparse and Low Rank Tensor Decomposition (NIPS 2015) with Parikshit Shah and Gongguo Tang
Collaborative Filtering with Graph Information: Consistency and Scalable Methods, (NIPS 2015) (spotlight) with Hsiang Fu Yu, Pradeep Ravikumar and Inderjit Dhillon [CODE]
PU Matrix Completion with Graph Information, (IEEE CAMSAP 2015) with Nagarajan Natarajan and Inderjit Dhillon
Forward - Backward Greedy Algorithms for Signal Demixing, (Asilomar, 2014) with Parikshit Shah and Stephen Wright
Sparse Overlapping Sets Lasso for Multitask Learning and its Application to fMRI Analysis, (NIPS 2013) (spotlight) with Chris Cox, Robert Nowak and Tim Rogers [CODE]
Adaptive Sensing with Structured Sparsity, (ICASSP 2013) with Gongguo Tang and Robert Nowak
A Greedy Forward-Backward Algorithm for Atomic Norm Constrained Minimization, (ICASSP 2013) with Parikshit Shah, Stephen Wright and Robert Nowak [CODE]
Correlated Gaussian Designs for Compressive Imaging, (IEEE ICIP, 2012) with Robert Nowak
A Clustering Approach to Optimize Online Dictionary Learning, (ICASSP 2012) with Fatih Porikli
Universal Measurement Bounds for Structured Sparse Signal Recovery (AISTATS 2012) with Ben Recht and Robert Nowak
Convex Approaches to Model Wavelet Sparsity Patterns, (IEEE ICIP 2011) (Best Student Paper Award) with Roberk Nowak, Steve Wright and Nick Kingsbury
Using Machines to Improve Human Saliency Detection, (Asilomar 2010) (oral) with Joe Harrison, Tyler Karrels, Robert Nowak and Timothy Rogers