Ankan Saha


I am a fifth year Graduate student in the department of Computer Science at the University of Chicago currently being advised by John Lafferty.
Partha Niyogi will always be the master.

I am broadly interested in topics related to the interplay between (Convex) Optimization and Machine Learning and their applications in Information Retrieval and Social Media.
I work on various areas of Machine learning including optimization of surrogate losses and online learning.
I also dabble with Learning Theory, Computational Geometry and Social Learning at times.

Please contact me at the email address below.

Contact
concatenation of first name and 's' [AT] cs [dot] uchicago [dot] edu


Work Experience

• Summers 2012: Research Intern in the Machine Learning group at MSR Redmond.
• Summers 2011: Research Intern at MSR Bangalore.
• Summers 2010: Research Intern at IBM Research Yorktown.
• Summers 2008: Research Intern at MPI Tubingen.

• Resume (.pdf)   

Publications

• Journals
  
  • On the Finite Time Convergence of Cyclic Coordinate Descent Methods
    with Ambuj Tewari
    Siam Journal of Optimization, 23(1):576-601, 2013. [.pdf]
    
    
  • Smoothing Multivariate Performance Measures
    with Xinhua Zhang and S.V.N. Vishwanathan
    Journal of Machine Learning Research (JMLR)
    Vol 13, pages 3589--3646, December, 2012. [.pdf] (Code)
    
    
  • Accelerated Training of Max-Margin Markov Networks with Kernels
    with Xinhua Zhang and S.V.N. Vishwanathan
    Journal of Theoretical Computer Science (TCS) (Accepted with Minor Revisions), 2012.
    
    
• Conferences
  
  • Learning Evolving and Emerging Topics in Social Media: A Dynamic NMF approach with Temporal Regularization
    with Vikas Sindhwani
    In the proceedings of WSDM 2012 [.pdf]   
    
    
  • Accelerated Training of Max-Margin Markov Networks with Kernels
    with Xinhua Zhang and S.V.N. Vishwanathan
    In the proceedings of ALT 2011 [.pdf]   
    
    
  • Smoothing Multivariate Performance Measures
    with Xinhua Zhang and S.V.N. Vishwanathan
    In the proceedings of UAI 2011 [.pdf]   
    
    
  • Improved Regret Guarantees for Online Smooth Convex Optimization with Bandit Feedback
    with Ambuj Tewari
    In the proceedings of AISTATS 2011(Oral Presentation)
    volume 15 of JMLR Workshop and Conference Proceedings, 2011 [.pdf] [html]    
    
    
  • Efficient Approximation Algorithms for Minimum Enclosing Convex Shapes
    with S.V.N. Vishwanathan and Xinhua Zhang
    In the proceedings of SODA 2011 [.pdf]   
    
    
  • On Data Independent Lower Bounds for SVM training
    with Xinhua Zhang and S.V.N. Vishwanathan
    In the proceedings of NIPS 2010 [.pdf]   
    
    
• Workshop Papers
  
  • Dynamic NMFs and Temporal Regularization for Online Analysis of Streaming Text
    with Vikas Sindhwani
    In the proceedings of NIPS 2010 workshop Machine Learning for Social Computing (Oral Presentation)[.pdf]    
    
    
• Preprints
  
  • The Interplay of Stability and Regret in Online Learning
    with Prateek Jain and Ambuj Tewari
    [.pdf]    
    
    
• Book Chapter
  
  • Semi-Supervised Learning: Some Recent Advances
    withXueyuan Zhou and Vikas Sindhwani
    in Cost-Sensitive Machine Learning   
    
    

Presentations

• Ankan Saha.(2012) Efficient Minimization of Risk Measures via Smoothing: Theory and Applications. Machine Learning Seminar University of California Santa Cruz, November 2012.
  
• Ankan Saha.(2012) Efficient Minimization of Risk Measures via Smoothing: Theory and Applications. Machine Learning Seminar, MSR Redmond, August 2012.
  
• Ankan Saha.(2012) Efficient Minimization of Risk Measures via Smoothing: Theory and Applications. Machine Learning Seminar, Google Research, August 2012.
  
• Ankan Saha.(2010) New Approximation Algorithms for Minimum Enclosing Convex Shapes. Machine Learning Seminar, Purdue University
  
• Ankan Saha.(2010) New Approximation Algorithms for Minimum Enclosing Convex Shapes. IBM TJ Watson Research Center, Yorktown,NY.
  
• Ankan Saha.(2010) Faster Rates for training Max-Margin Markov Networks. Machine Learning Reading Group, TTI Chicago.
  
  

Theses

• Topics in Structured Prediction: Problems and Approaches
  Masters Thesis, U.Chicago [Paper]

Teaching
Teaching Assistant for the following courses

• Winter 2010: Introduction to AI
  
• Summer 2010: Discrete Mathematics
  
• Spring 2009: Complexity Theory
  
• Winter 2009: Introduction to C++
  
• Fall 2008: Discrete Mathematics
  
• Winter 2007: Introduction to Computer Science
  
• Fall 2007: Algorithms
  

Fun Stuff
I used to organize the CS Grad Pizza Seminar at UChicago last year. Drop me an email if you want to give a talk.