Aaditya Ramdas Curriculum Vitae 1835 Cedar St, Apt B Berkeley, CA 94703 H (+1) 773-234-3277 B aramdas@cs.berkeley.edu Í cs.berkeley.edu/∼aramdas DOB: May 3, 1988 Research Interests My research spans theory, algorithms and applications in machine learning, nonparametric statistics and stochastic optimization. Broad areas of interest include { Sequential Learning { Kernel Methods { Nonparametric Hypothesis Testing { Statistical Learning Theory Specific recent subareas of interest within these include { Independence & Two-sample Testing { Statistics on Non-abelian Groups { Stochastic Convex Optimization { False Discovery Rate { Markov Chain Monte Carlo { Systems of Equations & Inequalities Current Position 2015–now Postdoctoral Researcher in EECS and Statistics. University of California, Berkeley (UCB), Berkeley (USA) Advisors: Michael Jordan & Martin Wainwright Academic Background 2013–15 Doctor of Philosophy in Statistics and Machine Learning. Carnegie Mellon University (CMU), Pittsburgh (USA) PhD Thesis: Computational and Statistical Advances in Testing and Learning Advisors: Larry Wasserman & Aarti Singh 2010–12 Master of Science in Machine Learning. Carnegie Mellon University (CMU), Pittsburgh (USA) GPA – 4.2 out of 4 Masters Data Analysis: Statistical Modeling of Crime in Pittsburgh 2005–09 Bachelor of Technology in Computer Science and Engineering. Indian Institute of Technology (IIT), Bombay (India) GPA – 9.44 out of 10 (Institute Rank 9/600, Department Rank 3/60) IIT Joint Entrance Exam, Rank 47/400,000. Work Experience 2009-10 Sum’14 Sum’12 Sum’09 Sum’08 Sum’07 Algorithmic Trader, Tower Research Capital, Gurgaon (India) & New York (USA). Research Intern (ML), Gatsby Neuroscience Unit (UCL), London (UK). Research Intern (AI), Microsoft Research (MSR), Cambridge (UK). Quantitative Analyst, Deutsche Bank, Mumbai (India). Research Intern (Geometry), INRIA, Sophia-Antipolis (France). Research Intern (Logic), LaBRI, Bordeaux (France). 1/7 Peer-Reviewed Publications ∗ = equal contribution, sub. = submitted, ar. = Arxiv preprint available (only full-length papers at competitive venues; excludes workshops and short papers at conferences with minimal refereeing) Hypothesis Testing sub., ar. Aaditya Ramdas, D. Isenberg∗ , A. Singh, L. Wasserman. Minimax Lower Bounds for Linear Independence Testing sub., ar. R. F. Barber, Aaditya Ramdas. The p-filter: multi-layer FDR control for grouped hypotheses ar. Aaditya Ramdas, A. Balsubramani∗ . Sequential Nonparametric Testing using the Law of the Iterated Logarithm sub., ar. Aaditya Ramdas, A. Singh, L. Wasserman. Classification Accuracy as a Proxy for Two Sample Testing ar. Aaditya Ramdas, N. Garcia∗ , M. Cuturi. Wasserstein Two Sample Testing and Related Families of Nonparametric Tests sub., ar. Aaditya Ramdas, S. Reddi, B. Poczos, A. Singh, L. Wasserman. Adaptivity & Computation-Statistics Tradeoffs for High Dimensional Two-Sample Testing NIPS K. Chwialkowski, Aaditya Ramdas, D. Sejdinovic, A. Gretton. Fast Two-Sample Testing with Analytic Representations of Probability Measures 29th Conference on Neural Information Processing Systems, 2015. IJCAI Aaditya Ramdas, L. Wehbe∗ . On the Increase in Power of Kernel Independence Testing due to Shrinkage 24th International Joint Conference on Artificial Intelligence, 2015. oral talk AISTATS Aaditya Ramdas, S. Reddi∗ , B. Poczos, A. Singh, L. Wasserman. High-Dimensional Power of Linear-Time Two Sample Tests vs. Mean-Shift Alternatives 18th International Conference on Artificial Intelligence & Statistics, 2015. AAAI Aaditya Ramdas, S. Reddi∗ , B. Poczos, A. Singh, L. Wasserman. On the Decreasing Power of Kernel & Distance based Hypothesis Tests in High Dimensions 29th AAAI Conference on Artificial Intelligence, 2015. Statistical Learning sub. M. Rabinovich, A. Ramdas, M. Wainwright, M. Jordan. Function mixing times and MCMC concentration away from equilibrium sub. A. El-Alaoui, X. Cheng, Aaditya Ramdas, M. Wainwright, M. Jordan. Asymptotic behavior of `q -based Laplacian regularization in semi-supervised learning sub. H. Mania, Aaditya Ramdas, M. Wainwright, M. Jordan, B. Recht. Universality of Mallows’ and degeneracy of Kendall’s kernels for rankings AISTATS Aaditya Ramdas, B. Poczos, A. Singh, L. Wasserman. An Analysis of Active Learning with Uniform Feature Noise 17th Intl. Conference on Artificial Intelligence & Statistics, 2014. oral talk ALT Aaditya Ramdas, A. Singh. Algorithmic Connections Between Active Learning and Stochastic Convex Optimization 24th International Conference on Algorithmic Learning Theory, 2013. oral talk 2/7 Convex Optimization sub., ar. Aaditya Ramdas, A. Hefny∗ , D. Needell∗ . Rows vs. Columns: Randomized Kaczmarz or Gauss-Seidel for Ridge Regression OMS Aaditya Ramdas, J. Peña. Towards a Deeper Geometrical, Analytical and Algorithmic Understanding of Margins Optimization Methods and Software, 2015. SIMAX Aaditya Ramdas, A. Ma∗ , D. Needell∗ . Convergence properties of the randomized extended Gauss-Seidel and Kaczmarz methods SIAM Journal of Matrix Analysis and Applications, 2015. ICML Aaditya Ramdas, J. Peña. Margins, Kernels and Non-linear Smoothed Perceptrons. 31st International Conference on Machine Learning, 2014. oral talk. ICML Aaditya Ramdas, A. Singh. Optimal Rates for Stochastic Convex Optimization under Tsybakov Noise Condition 30th International Conference on Machine Learning, 2013. oral talk Applications JCGS Aaditya Ramdas, R. Tibshirani. Fast & Flexible ADMM Algorithms for Trend Filtering Journal of Computational and Graphical Statistics, 2015. AoAS L. Wehbe, Aaditya Ramdas, R. Steorts, C. Shalizi. Regularized Brain Reading with Shrinkage and Smoothing Annals of Applied Statistics, 2015. PLoS ONE L. Wehbe, B. Murphy, P. Talukdar, A. Fyshe, Aaditya Ramdas, T. Mitchell. Simultaneously uncovering patterns of brain regions involved in story reading subprocesses Public Library of Science ONE, 2014. 3/7 Talks Conferences 2016 2015 2015 2015 2014 2014 2013 2013 2013 Beyond Worst Case Mixing Times for Markov Chains (ITA Workshop) Nonparametric Independence Testing for Small Sample Sizes (IJCAI) Adaptivity & Computation-Statistics Tradeoffs in Two Sample Testing (JSM) Sequential Testing using the Martingale LIL (Intl. W’shop on Sequential Methodology) Margins, Kernels and Nonlinear Smoothed Perceptrons (ICML) Active Learning with Uniform Feature Noise (AISTATS) Connecting Convex Optimization and Active Learning (NIPS, OPT Workshop) Algorithmic Connectiions between Convex Optimization and Active Learning (ALT) Optimal Convex Optimization under Tsybakov Noise Condition (ICML) Universities 2016 2015 2015 2014 2014 2014 2014 2014 2013 2013 2013 2013 p-Filter: FDR Control for Grouped Hypotheses (UC Davis) Insights into why Kernel Methods work well in High Dimensions (CMU) Margins - Algorithms, Geometry and Analysis (CMU) Computation-Statistics Tradeoffs in Two Sample Testing (UC Berkeley) Adaptivity for Two Sample Testing (Kyoto University) High Dimensional Two Sample Testing (Institute for Statistical Mathematics, Tachikawa) Algorithms for Trend Filtering (Gatsby Neuroscience Unit, University College London) Algorithms for Trend Filtering (CMU) Active Learning & Stochastic Optimization (Chennai Mathematical Institute) Active Learning & Stochastic Optimization (IIT Madras) Alg. Connections Between Active Learning and Stochastic Opt. (CMU) Optimal Upper and Lower Bounds for Stochastic Optimization (CMU) Industry 2016 2015 2015 2014 2013 2012 p-Filter: FDR Control for Grouped Hypotheses (AmpLab Retreat, Tahoe) Sequential Nonparametric Testing (Alibaba, Seattle) Sequential Nonparametric Testing (Google, Pittsburgh) Algorithms for Trend Filtering (Microsoft Research, Cambridge) Active Learning & Stochastic Optimization (IBM Research, Bangalore) Connecting Statistical & Logical Inference (Microsoft Research, Cambridge) Workshop Posters 2015 2015 2015 2015 2014 2014 2013 2013 NSF Workshop for Empirical Processes and Modern Statistical Decision Theory, Yale American Statistical Association (ASA), Pittsburgh Chapter, Best Poster Award UC San Diego, Information Theory & Applications (ITA) Workshop AIM Workshop on Inference in High Dimensional Regression UC Los Angeles, IPAM Stochastic Gradient Methods Workshop UC London, High-Dimensional Sensing and Inference Workshop GlobalSIP, IEEE Global Conference on Signal & Information Processing IISc Bangalore, Indo-US Machine Learning and Optimization Workshop 4/7 Professional Service Organization 2015 Workshop on Active Learning: Bridging Theory and Practice (ICML), Organizer. with Nina Balcan, Aarti Singh, Akshay Krishnamurthy. 2014 Workshop on Optimization in Machine Learning (NIPS), Organizer. with Alekh Agarwal, Suvrit Sra, Miro Dudik, Zaid Harchaoui, Martin Jaggi. 2014 ML Department Student Research Symposium (CMU), Lead Organizer. 2012-15 Weekly Lunch Seminar Series on Machine Learning (CMU), Organizer. Journal Reviewing 2016 2014,16 2015 2015 2015 2015 2014,15 2014 2014 2013 Annals of Applied Statistics (AoAS) Journal of Machine Learning Research (JMLR) Numerical Algorithms (NA) Machine Learning Journal (MACH) Journal of Artificial Intelligence Research (JAIR) Transactions on Pattern Analysis and Machine Intelligence (T-PAMI) Biometrika Data Mining and Knowledge Discovery (DAMI) Optimization Methods and Software (OMS) IEEE Transactions on Information Theory (IEEE-TIT) Conference Reviewing 2016 2015 2015 2014,15,16 2013,14 2013,16 2012 AAAI Conference on Artificial Intelligence (AAAI) European Conference on Machine Learning (ECML) International Joint Conference on Artificial Intelligence (IJCAI) International Conference on Machine Learning (ICML) Conference on Neural Information Processing Systems (NIPS) Conference on Artificial Intelligence & Statistics (AISTATS) Conference on Uncertainty in Artificial Intelligence (UAI) Department Service 2015-16 2014-15 2014-15 2013-14 2013-14 2008-09 2007-08 Graduate Admissions Committee, CS Department (UCB) Teaching Faculty Hiring Committee, ML Department (CMU) Graduate Student Assembly Representative (CMU), Outstanding Representative Award Graduate Admissions Committee, ML Department (CMU) Education Review Committee Founder, ML Department (CMU) Industry Job Placement Coordinator, CS Department (IITB) Sports Coordinator, CS Department (IITB) University Service 2015-16 2014-15 2014-15 2012-13 2011-12 2008-09 2008-09 2007-08 2007-08 2007-08 2006-07 Steward for Postdoctoral Union (UCB) SafeZone Allies for LGBTQ Safety, Trained Member (CMU) Campus Smoking Policy Review Committee, Member (CMU) Explorer’s Club Core Officer (CMU) Indian Graduate Students Association Treasurer (CMU) Campus Radio Cofounder (IITB) Job Fair Placement Representative, CS Department (IITB) Google Campus Ambassador (IITB) Institute Secretary for Academic Affairs (IITB) Sports Secretary, CS Department (IITB) Sports Secretary, Hostel 3 (IITB) 5/7 Teaching 2015 Alan J. Perlis Graduate Student Teaching Award, School of Computer Science. 2014 Best Teaching Assistant Award, Machine Learning Department. 2014 Machine Learning (MS), Review Videos, for all introductory graduate courses. 12 YouTube videos, 10-15 minutes each, on multivariate calculus, probability/statistics, real/functional analysis, linear algebra; review for the 400+ sized ML introductory course. 2013 Convex Optimization (PhD), Teaching Assistant, Ryan Tibshirani & Barnabas Póczós. Helped redesign course syllabus (content+schedule), restructured homeworks to increase flexibility, experimented with mastery questions, peer-grading, optional questions and project. 2012 Convex Optimization (PhD), Teaching Assistant, Geoff Gordon & Ryan Tibshirani. Created and graded homeworks and exams, gave recitations, held office hours, mentored course projects, maintained slides, scribes and videos. Guest Lectures (University) 2016 Sparse Linear Models, Statistical Learning Theory. (80 minutes class; course by Martin Wainwright) 2015 Active Learning (Distilled Sensing and Classification), Statistical Machine Learning. (80 minutes class; course by Larry Wasserman and Ryan Tibshirani) 2015 Modern Fast Stochastic Optimization for ML, Optimization. (80 minutes class; course by Ryan Tibshirani) 2014 Lower Bounds in Optimization, Advanced Optimization. (80 minutes class; course by Alex Smola and Suvrit Sra) 2012 ADMM and Mirror Descent, Convex Optimization. (80 minutes class; course by Geoff Gordon and Ryan Tibshirani) Outreach (Schools) 2016 Robots that run, Prescott Elementary, Oakland. (60 mins, once; special class for elementary students.) 2015 SVD, Random Graphs and Random Walks, PACT, Princeton. (90 mins, thrice; summer program for high school students.) 2015 Introduction to Computer Science, Technights, CMU. (90 min volunteering, twice; program to introduce middle school girls to computer science.) 2015 Introduction to Machine Learning, ISG, Muscat. (30 minutes class, twice; to introduce high school children to machine learning.) 2014 Mechanism Design: Auctions and Voting Theory, Andrew’s Leap, CMU. (80 minutes class; summer program for high school students) 2013 Online Learning: Multi-armed Bandits, Andrew’s Leap, CMU. (80 minutes class; summer program for high school students) Future Faculty Program Completed program run by the Eberly Center for Teaching Excellence (transcript available). 2014-15 Seminars. { { { { { Course & Syllabus Design Syllabus Design Workshop Promoting Peer Learning Planning & Delivering Effective Lectures Leveraging Diversity & Promoting Equity { { { { { Crafting a Teaching Statement Building a Teaching Portfolio Conducting Productive Discussions Engaging Students in Active Learning Good Assessment Practices 2014 Observations. { Classroom Teaching { Microteaching Workshop 2014 Projects. { Designed syllabus of UG course Mathematical Foundations of ML. { Pedagogical aspects of learning through videos. 6/7 Awards & Honors (includes a few that are both rarely awarded and competitive) 2013-14 Doug Beeferman PhD Fellowship, CMU. 2006-09 Inlaks Scholarship, IIT Bombay. Awarded to the single best all-round student out of all 600 undergraduates; given full tuition waiver plus additional stipend for three years. 2005-09 Academic Scholarships, CS Department and IIT Bombay. For securing top cumulative grade point average (CGPA), ending with department rank 3 (out of 60) and institute rank 9 (out of 600 undergrads). 2005 Guest of Indian Prime Minister’s Office. Invited to view India’s Annual Republic Day Parade from the PM’s Box, for academic excellence, including All India Rank 10 (out of 300,000) in Central Board’s Class 12 Final Exams with 97.4% and All India Rank 47 (out of 500,000) in the IIT Joint Entrance Exam. 2005-09 Institute Cultural Citation, IIT Bombay. Awarded to 5 out of 1000 graduating UG and PG students; for winning over 20 intercollegiate competitions in speaking, debating and literary arts, and several interhostel competitions. 2001-03 Sultanate of Oman, U-13 and U-15 Cricket Team. Represented Oman in the U-13 Gulf Cup, finishing runners-up, and in the U-15 Asia Cup, finishing 5th out of 14 teams. Also played for IIT Bombay and CMU cricket teams. 2001-03 Duke of Edinburgh Bronze and Silver Awards. Awarded the International Award for Young People under the Duke of Edinburgh Award Scheme, for social service, skill development, physical training and adventurous expeditions. 2013 Ironman Triathlon. Finishing medalist of the 3.8km swim, 180km bike, 42km run Ironman triathlon at Louisville, Kentucky (Aug 25, 2013). Also completed many other triathlons and marathons. 7/7