Sohail Bahmani

Postdoctoral Fellow
School of Electrical and Computer Engineering
Georgia Institute of Technology
Email: [first name]DOT[last name]AT[ece.gatech.edu]

Research Interest

My primary research interest is algorithmic and theoretical aspects of structured signal estimation in areas such as computational imaging, machine learning, and network analysis. My main motivation is to design provably accurate and efficient algorithms for structured inference problems that arise in different applications, using probability theory, information theory, statistics, and optimization.

Education and Appointments

  • Postdoc, Georgia Institute of Technology, 2013–present
  • PhD in ECE, Carnegie Mellon University, 2009–2013
  • Master's in Engineering Science, Simon Fraser University, Canada, 2007–2008
  • Bachelor's in EE, Sharif University of Technology, Iran, 2002–2006

Talks

  • Phase Retrieval Meets Statistical Learning Theory:
    IBM T.J. Watson Research Center, Apr. 2017
    Artificial Intelligence and Statistics conference (AISTATS'17), Apr. 2017
    Information Theory and Applications workshop (ITA'17), Feb. 2017
    Stochastic Seminar, School of Mathematics, Georgia Tech., Feb. 2017
  • Structured Matrix Estimation in High Dimensions
    School of Mathematics, University of Edinburgh, Jun. 2016

Publications

  • Journal Paper
  • Conference Paper
  • Preprint
  • Technical Report

Preprints

  1. S. Bahmani and J. Romberg, “Anchored Regression: Solving Random Convex Equations via Convex Programming,” submitted, Feb. 17, 2017. arXiv
  2. S. Bahmani, J. Romberg, P. Tetali, “Algebraic Connectivity Under Site Percolation in Finite Weighted Graphs,” in revision for IEEE Trans. on Network Science and Engineering, Sep. 2017. arXiv

2017

  1. S. Bahmani and J. Romberg, “Phase Retrieval Meets Statistical Learning Theory: A Flexible Convex Relaxation,” In Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS'17) , vol. 54 of Proceedings of Machine Learning Research , pp. 252–260. (Best paper award) arXivPMLR

2016

  1. S. Bahmani and J. Romberg, “Near-Optimal Estimation of Simultaneously Sparse and Low-Rank Matrices from Nested Linear Measurements,” Information and Inference, 5(3):331–351, 2016. arXivOxford Journals
  2. S. Bahmani, P. Boufounos, and B. Raj, “Learning Model-Based Sparsity via Projected Gradient Descent,” IEEE Trans. Info. Theory, 62(4):2092–2099, 2016. arXivIEEEXplore

2015

  1. S. Bahmani and J. Romberg, “Sketching for simultaneously sparse and low-rank covariance matrices,” in Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP'15), IEEE 6th International Workshop on, pp. 357–360, Cancun, Mexico, Dec. 2015.IEEEXplorearXiv
  2. S. Bahmani and J. Romberg, “Efficient Compressive Phase Retrieval with Constrained Sensing Vectors,” in Advances in Neural Information Processing Systems (NIPS'15), vol. 28, pp. 523–531, Montréal, Canada, Dec. 2015. arXivNIPS
  3. S. Bahmani and J. Romberg, “Lifting for Blind Deconvolution in Random Mask Imaging: Identifiability and Convex Relaxation,” SIAM Journal on Imaging Sciences, 8(4):2203–2238, 2015. arXiv SIAM
  4. S. Bahmani and J. Romberg, “Compressive Deconvolution in Random Mask Imaging,” IEEE Trans. on Computational Imaging, 1(4):236–246, 2015. arXivIEEEXplore

2013

  1. S. Bahmani, B. Raj, and P. T. Boufounos, “Greedy sparsity-constrained optimization,” Journal of Machine Learning Research, 14(3):807–841, 2013. JMLRarXivCode
  2. S. Bahmani, P. Boufounos, and B. Raj, “Robust 1-bit Compressive Sensing via Gradient Support Pursuit,” Apr. 2013. arXiv

2012

  1. S. Bahmani, B. Raj, “A unifying analysis of projected gradient descent for \(\ell_p\)-constrained least squares,” Applied and Computational Harmonic Analysis, 34(3):366–378, 2012. ElsevierarXiv

2011

  1. S. Bahmani, P. Boufonos, and B. Raj, “Greedy sparsity-constrained optimization,” in Conf. Record of the 45th Asilomar Conference on Signals, Systems, and Computers (ASILOMAR'11), pp. 1148–1152, Pacific Grove, CA, Nov. 2011. IEEEXploreSlidesCode

2010

  1. S. Bahmani, I. Bajić, and A. HajShirmohammadi, “Joint decoding of unequally protected JPEG2000 images and Reed-Solomon codes,” IEEE Trans. Image Processing, 19(10):2693–2704, Oct. 2010. IEEEXplore

2009

  1. S. Bahmani, I. Bajić, and A. HajShirmohammadi, “Improved Joint source channel decoding of JPEG2000 images and Reed-Solomon codes,” Proc. IEEE ICC'09, Dresden, Germany, Jun. 2009. IEEEXplore

2008

  1. S. Bahmani, I. Bajić, A. HajShirmohammadi, “Joint source channel decoding of JPEG2000 images with unequal loss protection,” Proc. IEEE ICASSP'08, pp. 1365–1368, Las Vegas, NV, Mar. 2008. IEEEXplore

Thesis

  • S. Bahmani, Algorithms for Sparsity-Constrained Optimization, PhD dissertation, Department of Electrical & Computer Engineernig, Carnegie Mellon University, Pittsburgh, PA, Feb. 2013. PDF