Aaditya Ramdas — aramdas [@] cs [dot] cmu [.] edu
8223 Gates (CMU), Pittsburgh

5th Year Joint PhD Student, Machine Learning and Statistics
School of Computer Science, Carnegie Mellon University

My advisors are Larry Wasserman (Stats) and Aarti Singh (ML).

I recently defended my thesis, and am moving to UC Berkeley as a postdoc in August, working with Martin Wainwright and Michael Jordan.

[Curriculum Vitae]


Research Interests

I actively work in the following research areas, along with some representative publications. Click here for the full list, along with coauthors, bibliography, links to talks, etc.

  1. Machine Learning (active learning)
    An Analysis of Active Learning with Uniform Feature Noise (AISTATS'14)
    Algorithmic Connections Between Active Learning and Stochastic Convex Optimization (ALT'13)

  2. Statistics (nonparametric hypothesis testing)
    On the High Dimensional Power of a Linear-Time Two Sample Test under Mean-shift Alternatives (AISTATS'15)
    Nonparametric Independence Testing for Small Sample Sizes (IJCAI'15)

  3. Optimization (for statistics/ML)
    Fast & Flexible ADMM Algorithms for Trend Filtering (JCGS'15)
    Margins, Kernels and Non-linear Smoothed Perceptrons (ICML'14)

  4. Optimization (general theory)
    Convergence properties of the randomized extended Gauss-Seidel and Kaczmarz methods (in submission)
    Towards A Deeper Geometric, Analytic and Algorithmic Understanding of Margins (in submission)

I also collaborate with neuroscientists, see the full list of such publications for more details.


Workshops, Seminars

In Jul. 2015, I co-organized (with Akshay Krishnamurthy, Nina Balcan and Aarti Singh) a workshop on Advances in Active Learning: Bridging Theory and Practice at ICML 2015.

In Dec. 2014, I co-organized (with Suvrit Sra, Zaid Harchaoui, Alekh Agarwal, Miro Dudik, Martin Jaggi) a workshop on Optimization for Machine Learning at NIPS 2014. [VIDEOS]

In Nov. 2014, I was the lead organizer (with great help from MLD students) of a full-day Machine Learning Department Student Research Symposium at CMU.

I also organized the CMU's weekly Machine Learning Lunch Seminar for over two years, and also a member of the Statistical ML Theory Group. Contact me if you want to speak at either!