Praneeth Netrapalli

My photo

Graduate Student
Dept. of Electrical and Computer Engineering
The University of Texas at Austin
Austin, TX - 78712, USA

E-mail: praneethn [at] utexas [dot] edu

Resume

About Me

I am a graduate student in the ECE department at UT Austin. My advisor is Prof. Sujay Sanghavi.

Education

Ph. D., Electrical and Computer Engineering, The University of Texas at Austin, August 2009 - present
M. S., Electrical and Computer Engineering, The University of Texas at Austin, August 2009 - May 2011
B. Tech., Electrical Engineering, Indian Institute of Technology Bombay, August 2003 - May 2007

Work Experience

Research Intern, Microsoft Research India, June 2012 - August 2012
Summer Intern, CNLS, Los Alamos National Laboratory, July 2010 - August 2010
Analyst, Core Quant Group, Goldman Sachs, Bangalore, June 2007 - July 2009

Research Interests

Machine Learning, Algorithms

Publications

  1. "Low-rank Matrix Completion using Alternating Minimization"
    P. Jain, P. Netrapalli and S. Sanghavi
    STOC 2013
    Alternating minimization (or alternating least squares a.k.a. ALS) is a popular heuristic to solve various low rank matrix problems. In this paper we provide the first theoretical guarantees for alternating minimization as applied to the low rank matrix completion problem.
  2. "Learning Markov Graphs Up To Edit Distance"
    A. K. Das, P. Netrapalli, S. Sanghavi and S. Vishwanath
    IEEE ISIT 2012
    In this paper, we derive lower bounds on the sample complexity for learning Ising and Gaussian graphical models up to an edit distance (Hamming distance). We use rate distortion approach to obtain strong converse results.
  3. "Finding the Graph of Epidemic Cascades"
    P. Netrapalli and S. Sanghavi
    ACM SIGMETRICS/Performance 2012
    We consider the inverse problem of learning the interaction graph by observing the spread of epidemics. For a popular epidemic spread model, called the Independent Cascade Model, we pose the maximum likelihood estimation as a convex optimization problem and derive theoretical guarantees on its performance. Applications include social networks, e-commerce etc.
  4. "Greedy Learning of Markov Network Structure"
    P. Netrapalli, S. Banerjee, S. Sanghavi and S. Shakkottai
    Allerton 2010 (Invited)
    We propose an efficient greedy algorithm to learn the structure of an undirected graphical model. We provide theoretical guarantees for the same under some (standard) assumptions for large girth graphs. We also demonstrate its efficiency empirically.
  5. "Learning Planar Ising Models"
    J. K. Johnson, P. Netrapalli and M. Chertkov
    Manuscript
    We propose a simple greedy algorithm for learning the best planar Ising model from a given set of samples. The efficiency of our algorithm relies on the fact that inference is tractable for the class of planar Ising models. We demonstrate the performance of our method empirically.

Talks

  1. "Low-rank Matrix Completion using Alternating Minimization"
    OPT workshop, NIPS 2012
    Slides Video
  2. "Learning the Graph of Epidemic Cascades"
    ACM SIGMETRICS/Performance 2012
    Slides