About Me

I am an Assistant Professor in the School of Computer Science in the College of Computing at the Georgia Institute of Technology.

My research focus is Machine Learning, and I like discovering connections between Optimization, Statistics, and Economics. I finished my PhD at UC Berkeley with Peter Bartlett in 2011, and I was a Simons postdoctoral fellow with Michael Kearns for the following two years.

Recent News

• ⇒ Thanks to the National Science Foundation for IIS Award 1910077 titled “RI: Small: Training Modularized Learning Systems Through Non-Convex Multiplayer Game Dynamics”
• ⇒ I am spending the summer of 2019 at Google Research NYC
• ⇒ Along with Shivani Agarwal, I will be the program co-chair of the 2020 Conference on Learning Theory (COLT)
• ⇒ In the years since Flint Michigan’s water crisis, I continue to work on the problem of detecting lead contamination and infrastructure remediation. More details on this project can be found here.

Current group

Former students and postdocs

Journal Papers

1. Abernethy, J., Lai, K. A., & Wibisono, A. (2019). Last-iterate convergence rates for min-max optimization. ArXiv Preprint ArXiv:1906.02027.
2. Abernethy, J., Lai, K. A., & Wibisono, A. (2019). Last-iterate convergence rates for min-max optimization. ArXiv Preprint ArXiv:1906.02027.
3. Misra, K., Schwartz, E. M., & Abernethy, J. (2018). Dynamic online pricing with incomplete information using multi-armed bandit experiments. Marketing Science (to Appear).
4. Abernethy, J., Lai, K. A., Levy, K. Y., & Wang, J.-K. (2018). Faster rates for convex-concave games. ArXiv Preprint ArXiv:1805.06792.
5. Chojnacki, A., Dai, C., Farahi, A., Shi, G., Webb, J., Zhang, D. T., … Schwartz, E. (2017). A Data Science Approach to Understanding Residential Water Contamination in Flint. ArXiv Preprint ArXiv:1707.01591.
6. Kodali, N., Abernethy, J., Hays, J., & Kira, Z. (2017). On convergence and stability of gans. ArXiv Preprint ArXiv:1705.07215.
7. Abernethy, J. D., Lee, C., McMillan, A., & Tewari, A. (2017). Online Learning via Differential Privacy. CoRR, abs/1711.10019. Retrieved from http://arxiv.org/abs/1711.10019
8. Abernethy, J., Anderson, C., Chojnacki, A., Dai, C., Dryden, J., Schwartz, E., … others. (2016). Data Science in Service of Performing Arts: Applying Machine Learning to Predicting Audience Preferences. ArXiv Preprint ArXiv:1611.05788.
9. Abernethy, J., Lee, C., & Tewari, A. (2016). Perturbation techniques in online learning and optimization. Perturbations, Optimization, and Statistics, Neural Information Processing Series.
10. Abernethy, J. D., Frongillo, R. M., & Kutty, S. (2015). On risk measures, market making, and exponential families. ACM SIGecom Exchanges, 13(2), 21–25.
11. Abernethy, J., Chen, Y., & Vaughan, J. W. (2013). Efficient market making via convex optimization, and a connection to online learning. ACM Transactions on Economics and Computation, 1(2), 12.
    PDF
12. Abernethy, J. D., Hazan, E., & Rakhlin, A. (2012). Interior-point methods for full-information and bandit online learning. IEEE Transactions on Information Theory, 58(7), 4164–4175.
    PDF
13. Abernethy, J. D. (2011). Sequential Decision Making in Non-stochastic Environments.
14. Abernethy, J., Chapelle, O., & Castillo, C. (2010). Graph regularization methods for web spam detection. Machine Learning Journal, 81(2), 207–225.
    PDF
15. Abernethy, J., Bach, F., Evgeniou, T., & Vert, J.-P. (2009). A New Approach to Collaborative Filtering: Operator Estimation with Spectral Regularization. The Journal of Machine Learning Research, 10, 803–826.
    PDF Abstract

    We present a general approach for collaborative filtering (CF) using spectral regularization to learn linear operators mapping a set of “users” to a set of possibly desired “objects”. In particular, several recent low-rank type matrix-completion methods for CF are shown to be special cases of our proposed framework. Unlike existing regularization-based CF, our approach can be used to incorporate additional information such as attributes of the users/objects—a feature currently lacking in existing regularization-based CF approaches—using popular and well-known kernel methods. We provide novel representer theorems that we use to develop new estimation methods. We then provide learning algorithms based on low-rank decompositions and test them on a standard CF data set. The experiments indicate the advantages of generalizing the existing regularization-based CF methods to incorporate related information about users and objects. Finally, we show that certain multi-task learning methods can be also seen as special cases of our proposed approach.
16. Abernethy, J., Evgeniou, T., Toubia, O., & Vert, J.-P. (2008). Eliciting consumer preferences using robust adaptive choice questionnaires. IEEE Transactions on Knowledge and Data Engineering, 20(2), 145–155.
    PDF

Conference and Workshop Papers

1. Cardoso, A. R., Abernethy, J., Wang, H., & Xu, H. (2019). Competing Against Nash Equilibria in Adversarially Changing Zero-Sum Games. In K. Chaudhuri & R. Salakhutdinov (Eds.), Proceedings of the 36th International Conference on Machine Learning (Vol. 97, pp. 921–930). Long Beach, California, USA: PMLR. Retrieved from http://proceedings.mlr.press/v97/cardoso19a.html
   Abstract

   We study the problem of repeated play in a zero-sum game in which the payoff matrix may change, in a possibly adversarial fashion, on each round; we call these Online Matrix Games. Finding the Nash Equilibrium (NE) of a two player zero-sum game is core to many problems in statistics, optimization, and economics, and for a fixed game matrix this can be easily reduced to solving a linear program. But when the payoff matrix evolves over time our goal is to find a sequential algorithm that can compete with, in a certain sense, the NE of the long-term-averaged payoff matrix. We design an algorithm with small NE regret–that is, we ensure that the long-term payoff of both players is close to minimax optimum in hindsight. Our algorithm achieves near-optimal dependence with respect to the number of rounds and depends poly-logarithmically on the number of available actions of the players. Additionally, we show that the naive reduction, where each player simply minimizes its own regret, fails to achieve the stated objective regardless of which algorithm is used. Lastly, we consider the so-called bandit setting, where the feedback is significantly limited, and we provide an algorithm with small NE regret using one-point estimates of each payoff matrix.
2. Wang, J.-K., & Abernethy, J. D. (2018). Acceleration through Optimistic No-Regret Dynamics. In Advances in Neural Information Processing Systems (pp. 3828–3838).
   PDF
3. Abernethy, J., Lai, K. A., Levy, K. Y., & Wang, J.-K. (2018). Faster Rates for Convex-Concave Games. In (COLT) Conference On Learning Theory (pp. 1595–1625).
   PDF
4. Abernethy, J., Chojnacki, A., Farahi, A., Schwartz, E., & Webb, J. (2018). ActiveRemediation: The Search for Lead Pipes in Flint, Michigan. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 5–14). ACM.
   PDF

   ** KDD 2018 Best Student Paper Award (Research Track) **
5. Abernethy, J. D., & Wang, J.-K. (2017). On frank-wolfe and equilibrium computation. In Advances in Neural Information Processing Systems (pp. 6584–6593).
   PDF
6. Chojnacki, A., Dai, C., Farahi, A., Shi, G., Webb, J., Zhang, D. T., … Schwartz, E. (2017). A Data Science Approach to Understanding Residential Water Contamination in Flint. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1407–1416). ACM.
   PDF
7. Abernethy, J., & Hazan, E. (2016). Faster Convex Optimization: Simulated Annealing with an Efficient Universal Barrier. In International Conference on Machine Learning (pp. 2520–2528).
   PDF
8. Abernethy, J., Lahaie, S., & Telgarsky, M. (2016). Rate of Price Discovery in Iterative Combinatorial Auctions. In Proceedings of the 2016 ACM Conference on Economics and Computation (pp. 809–809). ACM.
   PDF
9. Azab, M., Mihalcea, R., & Abernethy, J. (2016). Analysing RateMyProfessors Evaluations Across Institutions, Disciplines, and Cultures: The Tell-Tale Signs of a Good Professor. In International Conference on Social Informatics (pp. 438–453). Springer International Publishing.
   PDF
10. Li, J., Ozog, P., Abernethy, J., Eustice, R. M., & Johnson-Roberson, M. (2016). Utilizing high-dimensional features for real-time robotic applications: Reducing the curse of dimensionality for recursive Bayesian estimation. In Intelligent Robots and Systems (IROS), 2016 IEEE/RSJ International Conference on (pp. 1230–1237). IEEE.
    PDF
11. Abernethy, J. D., Amin, K., & Zhu, R. (2016). Threshold bandits, with and without censored feedback. In Advances in Neural Information Processing Systems (To Appear) (pp. 4889–4897).
    PDF
12. Abernethy, J., & Johnson-Roberson, M. (2015). Financialized methods for market-based multi-sensor fusion. In IEEE/RSJ International Conference on Intelligent Robots and Systems.
    PDF
13. Abernethy, J., Lahaie, S., & Telgarsky, M. (2015). Price Discovery in Subgradient Combinatorial Auctions. In Auctions, Market Mechanisms, and Their Applications. AMMA Non-archival Track.
    PDF
14. Waggoner, B., Frongillo, R., & Abernethy, J. (2015). A Market Framework for Eliciting Private Data. In Advances in Neural Information Processing Systems 28.
    PDF
15. Abernethy, J., Lee, C., & Tewari, A. (2015). Fighting Bandits with a New Kind of Smoothness. In Advances in Neural Information Processing Systems 28.
    PDF
16. Abernethy, J., Chen, Y., Ho, C.-J., & Waggoner, B. (2015). Low-cost learning via active data procurement. In Proceedings of the Sixteenth ACM Conference on Economics and Computation (pp. 619–636). ACM.
    PDF
17. Abernethy, J., Kutty, S., Lahaie, S., & Sami, R. (2014). Information aggregation in exponential family markets. In (EC) Proceedings of the 15th ACM conference on Economics and computation (pp. 395–412). ACM.
    PDF
18. Wang, Q., Sheng, S.-P., Abernethy, J., & Liu, M. (2014). Jamming Defense Against a Resource-Replenishing Adversary in Multi-channel Wireless Systems. In (WiOpt) International Symposium on Modeling and Optimization in Mobile, Ad Hoc and Wireless Networks.
    PDF
19. Abernethy, J. D., Frongillo, R. M., Li, X., & Vaughan, J. W. (2014). A general volume-parameterized market making framework. In (EC) Proceedings of the 15th ACM conference on Economics and computation (pp. 413–430). ACM.
    PDF
20. Abernethy, J., Lee, C., Sinha, A., & Tewari, A. (2014). Online Linear Optimization via Smoothing. In (COLT) Conference on Learning Theory.
    PDF
21. Abernethy, J., Amin, K., Draief, M., & Kearns, M. (2013). Large-Scale Bandit Problems and {KWIK} Learning. In (ICML) Proceedings of the 30th International Conference on Machine Learning (pp. 588–596).
    PDF
22. McMahan, B., & Abernethy, J. (2013). Minimax optimal algorithms for unconstrained linear optimization. In (NIPS) Advances in Neural Information Processing Systems (pp. 2724–2732).
    PDF
23. Abernethy, J., & Kale, S. (2013). Adaptive Market Making via Online Learning. In (NIPS) Advances in Neural Information Processing Systems (pp. 2058–2066).
    PDF
24. Abernethy, J., Bartlett, P., Frongillo, R., & Wibisono, A. (2013). How to Hedge an Option Against an Adversary: Black-Scholes Pricing is Minimax Optimal. In (NIPS) Advances in Neural Information Processing Systems (pp. 2346–2354).
    PDF
25. Abernethy, J., Frongillo, R. M., & Wibisono, A. (2012). Minimax option pricing meets black-scholes in the limit. In (STOC) Proceedings of the 44th Symposium on Theory of Computing (pp. 1029–1040). ACM.
    PDF
26. Abernethy, J. D., & Frongillo, R. M. (2012). A Characterization of Scoring Rules for Linear Properties. In (COLT) Conference on Learning Theory.
    PDF
27. Abernethy, J., & Frongillo, R. M. (2011). A Collaborative Mechanism for Crowdsourcing Prediction Problems. In (NIPS) Advances in Neural Information Processing Systems (pp. 2600–2608).
    PDF
28. Abernethy, J., Chen, Y., & Wortman Vaughan, J. (2011). An optimization-based framework for automated market-making. In (EC) Proceedings of the 12th ACM conference on Electronic commerce (pp. 297–306). ACM.
    PDF
29. Abernethy, J., Bartlett, P. L., & Hazan, E. (2011). Blackwell approachability and low-regret learning are equivalent. In (COLT) Conference on Learning Theory.
    PDF
30. Abernethy, J., & Warmuth, M. K. (2010). Repeated Games against Budgeted Adversaries. In (NIPS) Advances in Neural Information Processing Systems (pp. 1–9).
    PDF
31. Abernethy, J., Bartlett, P. L., Buchbinder, N., & Stanton, I. (2010). A regularization approach to metrical task systems. In (ALT) Algorithmic Learning Theory (pp. 270–284). Springer Berlin Heidelberg.
    PDF
32. Abernethy, J., & Rakhlin, A. (2009). Beating the adaptive bandit with high probability. In (COLT) Conference on Learning Theory.
    PDF
33. Abernethy, J., Agarwal, A., Bartlett, P. L., & Rakhlin, A. (2009). A stochastic view of optimal regret through minimax duality. In (COLT) Conference on Learning Theory.
    PDF
34. Abernethy, J., Chapelle, O., & Castillo, C. (2008). Web spam identification through content and hyperlinks. In (AirWeb) Proceedings of the 4th international workshop on Adversarial information retrieval on the web (pp. 41–44). ACM.
    PDF
35. Abernethy, J., Bartlett, P. L., Rakhlin, A., & Tewari, A. (2008). Optimal strategies and minimax lower bounds for online convex games. In (COLT) Conference on Learning Theory.
    PDF
36. Abernethy, J., Hazan, E., & Rakhlin, A. (2008). Competing in the Dark: An Efficient Algorithm for Bandit Linear Optimization. In (COLT) Conference on Learning Theory.
    PDF

    ** COLT 2008 Best Student Paper Award & 2008 Pat Goldberg Memorial Best Paper Award from IBM **
37. Abernethy, J., Warmuth, M. K., & Yellin, J. (2008). Optimal strategies from random walks. In (COLT) Conference on Learning Theory.
    PDF
38. Rakhlin, A., Abernethy, J., & Bartlett, P. L. (2007). Online discovery of similarity mappings. In (ICML) Proceedings of the 24th international conference on Machine learning (pp. 767–774). ACM.
    PDF
39. Abernethy, J., Bartlett, P., & Rakhlin, A. (2007). Multitask learning with expert advice. In (COLT) Conference on Learning Theory.
    PDF
40. Abernethy, J., Langford, J., & Warmuth, M. K. (2006). Continuous experts and the binning algorithm. In (COLT) Conference on Learning Theory.
    PDF

Technical Report

1. Waggoner, B., Frongillo, R. M., & Abernethy, J. D. (2017). Addendum to "A Market Framework for Eliciting Private Data". CoRR (Vol. abs/1703.00899). Retrieved from http://arxiv.org/abs/1703.00899
2. Kodali, N., Abernethy, J. D., Hays, J., & Kira, Z. (2017). How to Train Your DRAGAN. CoRR (Vol. abs/1705.07215). Retrieved from http://arxiv.org/abs/1705.07215
3. Abernethy, J. D., Anderson, C., Dai, C., Farahi, A., Nguyen, L., Rauh, A., … Yang, S. (2016). Flint Water Crisis: Data-Driven Risk Assessment Via Residential Water Testing. CoRR (Vol. abs/1610.00580). Retrieved from http://arxiv.org/abs/1610.00580
4. Abernethy, J., & Hazan, E. (2015). Faster Convex Optimization: Simulated Annealing with an Efficient Universal Barrier. CoRR (Vol. abs/1507.02528). Retrieved from http://arxiv.org/abs/1507.02528
5. Abernethy, J. D., Lee, C., & Tewari, A. (2015). Spectral Smoothing via Random Matrix Perturbations. CoRR (Vol. abs/1507.03032). Retrieved from http://arxiv.org/abs/1507.03032
6. Abernethy, J. D., Chen, Y., Ho, C.-J., & Waggoner, B. (2015). Actively Purchasing Data for Learning. CoRR (Vol. abs/1502.05774). Retrieved from http://arxiv.org/abs/1502.05774
7. Abernethy, J., Kutty, S., Lahaie, S., & Sami, R. (2014). Information Aggregation in Exponential Family Markets. CoRR (Vol. abs/1402.5458).
8. Abernethy, J., Lee, C., Sinha, A., & Tewari, A. (2014). Online Linear Optimization via Smoothing. CoRR (Vol. abs/1405.6076).
9. Abernethy, J., Frongillo, R. M., & Wibisono, A. (2012). Minimax Option Pricing Meets Black-Scholes in the Limit. CoRR (Vol. abs/1202.2585).
10. Abernethy, J., & Frongillo, R. M. (2011). A Collaborative Mechanism for Crowdsourcing Prediction Problems. CoRR (Vol. abs/1111.2664).
11. Abernethy, J., Chen, Y., & Vaughan, J. W. (2010). An Optimization-Based Framework for Automated Market-Making. CoRR (Vol. abs/1011.1941).
12. Abernethy, J., Bartlett, P. L., & Hazan, E. (2010). Blackwell Approachability and Low-Regret Learning are Equivalent. CoRR (Vol. abs/1011.1936).
13. Abernethy, J., Agarwal, A., Bartlett, P. L., & Rakhlin, A. (2009). A Stochastic View of Optimal Regret through Minimax Duality. CoRR (Vol. abs/0903.5328).
14. Bach, F., Abernethy, J., Vert, J.-P., & Evgeniou, T. (2008). A New Approach to Collaborative Filtering: Operator Estimation with Spectral Regularization. CoRR (Vol. abs/0802.1430).
15. Abernethy, J., Bach, F., Evgeniou, T., & Vert, J.-P. (2006). Low-rank matrix factorization with attributes. CoRR (Vol. abs/cs/0611124).

Open Problems

1. Abernethy, J., & Mannor, S. (2011). Does an Efficient Calibrated Forecasting Strategy Exist? (COLT) Conference on Learning Theory.
   PDF
2. Abernethy, J. (2010). Can We Learn to Gamble Efficiently? (COLT) Conference on Learning Theory.
   PDF
3. Abernethy, J., & Rakhlin, A. (2009). An Efficient Bandit Algorithm for root-T-Regret in Online Multiclass Prediction? (COLT) Conference on Learning Theory.
   PDF
4. Abernethy, J., & Warmuth, M. K. (2009). Minimax Games with Bandits. (COLT) Conference on Learning Theory.
   PDF

Talks

Fun Facts

Courses I have taught

Number	Name	Intitution	Year
CS 7505	Machine Learning Theory	Georgia Tech	Fall 2018	VIEW
CS 4540	Advanced Algorithms for Machine Learning	Georgia Tech	Fall 2018	VIEW
CS 7505	Machine Learning Theory	Georgia Tech	Fall 2017
EECS 445	Machine Learning	Univ. Michigan	Fall 2016	VIEW
EECS 545	Graduate Machine Learning	Univ. Michigan	Winter 2016	VIEW
EECS 598	Theoretical Foundations of Machine Learning	Univ. Michigan	Fall 2015	VIEW
EECS 203	Introduction to Discrete Mathematics	Univ. Michigan	Fall 2014	VIEW
EECS 203	Introduction to Discrete Mathematics	Univ. Michigan	Winter 2014	VIEW
EECS 598	Prediction and Learning: It's Only a Game	Univ. Michigan	Fall 2013	VIEW