Welcome to my site.

Mesrob I. Ohannessian

Mesrob Ohannessian
Assistant Professor
Electrical and Computer Engineering
University of Illinois at Chicago
[first name]@uic.edu
 

Previously: Research Assistant Professor at the Toyota Technological Institute at Chicago (TTIC), postdoc at UC San Diego with Alon Orlitsky, visiting scholar at the Simons Institute at UC Berkeley, postdoc at MSR-Inria with Laurent Massoulié, ERCIM Marie Curie Fellow at the Université Paris-Sud with Stéphane Boucheron and Elisabeth Gassiat

Education: PhD from MIT, with Munther Dahleh and Sanjoy Mitter, at the Laboratory for Information and Decision Systems (LIDS)


Google Scholar
  |   Publications   |   Teaching



Research


My research interests are broadly in machine learning, statistics, and information theory, and their applications. My goal is to draw the most benefit from the smallest amount of data while addressing the statistical, computational, and societal aspects of modern data analysis.
I am currently interested in two main problems: designing learning algorithms that adapt to structure in data and making non-discriminatory algorithmic decisions both computationally feasible and aware of their long-term societal impact.

Recent News

- August 2020: "Towards competitive N-gram smoothing", with M. Falahatgar, A. Orlitsky, and V. Pichapati, will appear in AISTATS 2020!
- July 2020: "Fair Learning with Private Demographic Data", with H. Mouzannar and N. Srebro, will appear in ICML 2020!
- Spring 2020: I taught a new course: ECE 491 Information and Learning.
- December 2019: We had an oral presentation on "Fair Learning with Private Demographic Data", with H. Mouzannar and N. Srebro, at the NeurIPS 2019 Workshop on ML with Guarantees.
- Fall 2019: I taught ECE 341, an introductory probability course for engineers.
- August 2019: I joined the Department of Electrical and Computer Engineering at UIC!
- August 2019: I was invited and gave a talk at Elucid8 at UW Madison
- July-August 2019: I taught (for the second time!) "How Machines Learn" at Summer Lab, to rising 7th and 8th graders.
- April 2019: I was invited and gave talks on "From Fair Decisions to Social Benefit" at the ML Seminar at the University of Chicago and the SINE Seminar at UIUC
- October 2018: "From Fair Decision Making to Social Equality", with H. Mouzannar and N. Srebro, accepted at ACM FAT* 2019.
- June-July 2018: I developed and taught a three-week activity-based course introducing machine learning and Python coding to rising 7th and 8th grade students at the University of Chicago Lab Schools' Summer Lab program.
- June 2018: I gave an invited talk titled "One (categorical) distribution estimator for all dimensions" at the 2018 Midwest Machine Learning Symposium.




Publications

Statistical Learning and Information Theory

"Towards competitive N-gram smoothing". M. Falahatgar, —, A. Orlitsky, and V. Pichapati. AISTATS. August 26-28, 2020. 

"Fair Learning with Private Demographic Data". H. Mozannar, —, N. Srebro.  ICML. July 12-18, 2020.

"From Fair Decision Making to Social Equality". H. Mozannar, —, N. Srebro. ACM FAT*, 2019.

"On the Impossibility of Learning the Missing Mass". E. Mossel and —. Entropy. 21(1), 28, 2019.

"A Truncation Model for Estimating Species Richness". F. Koladjo, —, E. Gassiat. International Journal of Biostatistics (special issue).Published online, July 2018.

"The Power of Absolute Discounting - All-Dimensional Distribution Estimation". M. Falahatgar, —, A. Orlitsky, and V. Pichapati. NIPS. (Long Beach, CA) December 5-9, 2017.

"Learning Non-Discriminatory Predictors". B. Woodworth, S. Gunasekar, —, and N. Srebro. 30th Ann. Conf. on Learning Theory (COLT). (Amsterdam, Netherlands) July 7-10, 2017.

"Concentration Inequalities in the Infinite Urn Scheme for Occupancy Counts and the Missing Mass, with Applications". A. Ben-Hamou, S. Boucheron, and —.  Bernoulli, 23(1):249-287, 2017. Preliminary poster "Estimating the Small Data in Big Data" appeared at Analysis of Algorithms (AofA) 2014.

Near-Optimal Smoothing of Structured Conditional Probability Matrices”. M. Falahatgar, —, and A. Orlitsky. NIPS. (Barcelona, Spain) December 5-10, 2016.

"Greedy-Bayes for Targeted News Dissemination". L. Massoulié, —, and A. Proutière. ACM SIGMETRICS. (Portland, Oregon) June 15-19, 2015.

"Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning". M. Lučić, —, A. Karbasi, and A. Krause. 18th International Conference on Artificial Intelligence and Statistics (AISTATS). (San Diego, California) May 9-12, 2015. Best student paper award.

"About Adaptive Coding on Countable Alphabets: Max-Stable Envelope Classes". S. Boucheron, E. Gassiat, and —. IEEE Transactions on Information Theory, 61(9):4948–4967, 2015.

Rare Probability Estimation under Regularly Varying Heavy Tails”. — and M. A. Dahleh. 25th Ann. Conf. on Learning Theory (COLT). (Edinburgh, Scotland) June 25-27, 2012.

Large Alphabets: Finite, Infinite, and Scaling Models”. — and M. A. Dahleh. 46th Ann. Conf. on Information Sciences and Systems (CISS). (Princeton, NJ) March 21-23, 2012. (Invited)

"On Inference about Rare Events". PhD Thesis. MIT, Cambridge, MA, USA. February 2012.

Canonical Estimation in a Rare-Events Regime”. —, V. Y. F. Tan and M. A. Dahleh. Allerton Conference. (Monticello, IL) September 28-30, 2011.

Distribution-Dependent Performance of the Good-Turing Estimator for the Missing Mass”. — and M. A. Dahleh. 19th Int. Symp. on Mathematical Theory of Networks and Systems (MTNS). (Budapest, Hungary) July 5-9, 2010.

Structure Learning in Causal Cyclic Networks”. S. Itani, —, K. Sachs, G. P. Nolan, and M. A. Dahleh. Proc. of NIPS 2008 Workshop on Causality, Journal of Machine Learning Research (NIPS Workshop), 6:165-176, 2010.

A Turbo-Style Algorithm for Lexical Baseforms Estimation”. G. F. Choueiter, —, S. Seneff, and J. R. Glass. IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP). (Las Vegas, NV) March 30-April 4, 2008.
 

Optimization and Control

"Robust and Optimal Consumption Policies for Deadline-Constrained Deferrable Loads". M. Roozbehani, D. Materassi, —, and M. A. Dahleh. IEEE Transactions on Smart Grids, 5(4):1823–1834, 2014.

"Dynamic Estimation of Price-Response of Deadline-Constrained Electric Loads with Threshold Policies". —, M. Roozbehani, D. Materassi and M. A. Dahleh. American Control Conference (ACC). 2014.

The Intertemporal Utility of Demand and Price-Elasticity of Consumption in Power Grids with Shiftable Loads”. M. Roozbehani, A. Faghih, — and M. A. Dahleh. 50th IEEE Conf. on Decision and Control (CDC). (Orlando, FL) December 12-15, 2011.

Optimal Caching Router Placement for Reduction in Retransmission Delay”. M. P. McGarry, R. Shakya, — and R. Ferzli. 20th Int. Conf. on Computer Communications and Networks (ICCCN). (Lahaina, HI) July 31-August 4, 2011.


Signal Processing and Other Topics

 “A Pilot Project – From Illiteracy to Computer Literacy: Teaching and Learning Using Information Technology”. M. A. Al-Alaoui, —, G. F. Choueiter, C. Akl, T. T. Avakian, I. Al-Kamal, and R. Ferzli. Int. Journal of Emerging Technologies in Learning, 3(3):4-9, 2008. Preliminary version in ICL 2007. (Villach, Austria) September 26-28, 2007.

“Complex-Coefficient Polynomial Roots by a Stability Criterion”. —, K. Kabalan and A. El-Hajj. Int. Journal of Information and Systems Sciences, 1(1): 89–104, 2005.

“A Novel Parallel Architecture for Local Histogram Equalization”. —, G. F. Choueiter, H. Diab. SPIE Visual Communications and Image Processing (VCIP). (Beijing, China) July 12-15, 2005.

"Simulation and Visualization of Fields and Energy Flows in Electric Circuits with Idealized Geometries". SM Thesis. MIT, Cambridge, MA, USA. June 2005.

“Speech Recognition for Voice Based Control”. Y. S. Naous, G. F. Choueiter, — and M. A. Al-Alaoui. IEEE Symposium on Signal Processing and Information Technology (ISSPIT). (Marrakesh, Morocco) December 18-21, 2002.



Recent Teaching

UIC

SP 2020:   ECE 491. Information and Learning
FA 2019:   ECE 341. Probability and Stochastic Processes for Engineers


Earlier 

Summer 2018 and Summer 2019: "How Machines Learn" at Summer Lab at the University of Chicago Lab Schools 
(course for rising 7th and 8th grade students)
Fall Quarter 2017: TTIC 31150/CMSC 31150 Mathematical Toolkit at TTIC and the University of Chicago




Address: Mesrob I. Ohannessian
                 Electrical and Computer Engineering
                 University of Illinois at Chicago
                 851 S. Morgan St., 1030 SEO (MC 154)
                 Chicago,IL 60607-7042