Peter Lofgren
 |
I'm excited about data-mining, efficient infrastructure, and machine learning. For my PhD I developed more efficient algorithms for Personalized PageRank, a model of user interests on networks which is used for friend recommendation and personalized search. My awesome PhD advisor was Ashish Goel. I've interned at Google, LinkedIn, and a start-up. After graduating, I developed a high performance graph library and applied deep learning to photos at Teapot, and now I'm using machine learning to thwart credit card fraud at Stripe.
Fun fact: I appear in the Hollywood movie The Internship which was filmed at Google while I was an intern there. 20 second clip
Publications
PageRank and Random Walks
Peter Lofgren (joint work with Siddhartha Banerjee and Ashish Goel): Efficient Algorithms for Personalized PageRank. PhD Thesis.
Presents our bidirectional algorithms and the prior algorithms they depend on.
Peter Lofgren, Siddhartha Banerjee, and Ashish Goel: Personalized PageRank Estimation and Search: A Bidirectional Approach. WSDM 2016
This bidirectional estimator is more efficient and simpler to analyze than our previous estimator, FAST-PPR.
I've implemented our algorithm (with unit tests) in scala at GitHub and in C++ as part of SNAP.
Siddhartha Banerjee, Peter Lofgren: Fast Bidirectional Probability Estimation in Markov Models. NIPS 2015
This generalizes our bidirectional estimator to arbitrary Markov Chains, and allows fast estimation of the Heat Kernel between a pair of nodes. (Experiment Code)
Peter Lofgren, Siddhartha Banerjee, Ashish Goel: Bidirectional PageRank Estimation: From Average-Case to Worst-Case. Workshop on Algorithms and Models for the Web Graph (WAW) 2015.
This gives an alternative bidirectional estimator for Personalized PageRank on undirected graphs.
Peter Lofgren, Siddhartha Banerjee, Ashish Goel, and C. Seshadhri: FAST-PPR: Scaling Personalized PageRank Estimation for Large Graphs. KDD 2014
Open source code is available at github.
KDD talk (12 min) and slides (pptx) (pdf)
Peter Lofgren, Ashish Goel: Personalized PageRank to a Target Node. Technical Report 2013
Peter Lofgren: On the complexity of the Monte Carlo method for incremental PageRank. Information Processing Letters 114(3): 104-106 (2014)
Crowdsourcing
Steven Euijong Whang, Peter Lofgren, Hector Garcia-Molina: Question Selection for Crowd Entity Resolution. PVLDB 6(6): 349-360 (2013)
Vasilis Verroios, Peter Lofgren, and Hector Garcia-Molina: tDP: An Optimal-Latency Budget Allocation Strategy for Crowdsourced MAXIMUM Operations. SIGMOD 2015
Undergraduate work on Cryptographic Privacy
Peter Lofgren, Nicholas Hopper: BNymble: More Anonymous Blacklisting at Almost No Cost (A Short Paper). Financial Cryptography 2011: 268-275
Peter Lofgren, Nicholas Hopper: FAUST: efficient, TTP-free abuse prevention by anonymous whitelisting. WPES 2011: 125-130
Undergraduate Work on Knot Theory
E. Bunch, P. Lofgren, A. Rapp and D. N. Yetter: On quotients of quandles. Journal of Knot Theory and Its Ramifications: 19(09), 1145-1156 (2010)