Publications.

CapuchinAI 1.0: Development of a machine learning-based touchscreen paradigm to test cognition in wild capuchins

Sánchez Vargas, F., Potluri, S. R., Abernethy, J. D., & Benítez, M. E.

American Journal of Primatology. Published online July 28, 2026. DOI: 10.1002/ajp.70194.

Multi-distribution learning: from worst-case optimality to lexicographic min-max optimality

Wang, G., Syed, U., Schapire, R. E., & Abernethy, J. D.

37th International Conference on Algorithmic Learning Theory (ALT 2026). Accepted.

Faster margin maximization rates for generic and adversarially robust optimization methods

Wang, G., Hu, Z., Gentile, C., Muthukumar, V., & Abernethy, J. D.

Mathematical Programming. Published online October 9, 2025. DOI: 10.1007/s10107-025-02283-4.

Can transformers reason logically? A study in SAT solving

Pan, L., Ganesh, V., Abernethy, J. D., Esposo, C., & Lee, W.

Proceedings of the 42nd International Conference on Machine Learning, PMLR 267, 47632–47671.

Symbolic constraint-solving capabilities of transformer large language models

Pan, L., Esposo, C., Abernethy, J. D., Ganesh, V., & Lee, W.

Proceedings of the 1st GENZERO Workshop: Revolutionizing Autonomous Systems with Generative AI, 147–154. DOI: 10.1007/978-981-95-1050-4_18.

SPARQ: Synthetic problem generation for reasoning via quality-diversity algorithms

Havrilla, A., Hughes, E., Samvelyan, M., & Abernethy, J. D.

arXiv preprint arXiv:2506.06499.

Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities

Google DeepMind Gemini Team (including Abernethy, J.).

arXiv technical report arXiv:2507.06261.

No-regret dynamics in the Fenchel game: a unified framework for algorithmic convex optimization

Wang, J.-K., Abernethy, J. D., & Levy, K. Y.

Mathematical Programming, 205 (1–2), 203–268.

Extragradient type methods for Riemannian variational inequality problems

Hu, Z., Wang, G., Wang, X., Wibisono, A., Abernethy, J. D., & Tao, M.

Proceedings of AISTATS 2024, PMLR 238, 2080–2088.

Lexicographic optimization: algorithms and stability

Abernethy, J. D., Schapire, R. E., & Syed, U.

Proceedings of AISTATS 2024, PMLR 238, 2503–2511.

A mechanism for sample-efficient in-context learning for sparse retrieval tasks

Abernethy, J. D., Agarwal, A., Marinov, T. V., & Warmuth, M. K.

Proceedings of ALT 2024, PMLR 237, 3–46.

Bring human values to AI

Abernethy, J. D., Candelon, F., Evgeniou, T., Gupta, A., & Lostanlen, Y.

Harvard Business Review, March–April 2024.

Accelerated federated optimization with quantization

Youn, Y., Kumar, B., & Abernethy, J. D.

IEEE Data Engineering Bulletin, 46 (1), 79–123.

Artificial intelligence for climate smart forestry: a forward looking vision

Luo, F., Liu, L., Wang, G. G., Kumar, V., Ashton, M. S., Abernethy, J. D., et al.

Proceedings of IEEE CogMI 2023, 1–10. DOI: 10.1109/CogMI58952.2023.00011.

Minimizing dynamic regret on geodesic metric spaces

Hu, Z., Wang, G., & Abernethy, J. D.

Proceedings of COLT 2023, PMLR 195, 4336–4383.

On accelerated perceptrons and beyond

Wang, G., Hanashiro, R., Guha, E. K., & Abernethy, J. D.

International Conference on Learning Representations (ICLR 2023).

Riemannian projection-free online learning

Hu, Z., Wang, G., & Abernethy, J. D.

Advances in Neural Information Processing Systems 36, 41980–42014.

Faster margin maximization rates for generic optimization methods

Wang, G., Hu, Z., Muthukumar, V., & Abernethy, J. D.

Advances in Neural Information Processing Systems 36, 62488–62518.

Active sampling for min-max fairness

Abernethy, J. D., Awasthi, P., Kleindessner, M., Morgenstern, J., Russell, C., & Zhang, J.

Proceedings of ICML 2022, PMLR 162, 53–65.

ActiveHedge: Hedge meets active learning

Kumar, B., Abernethy, J. D., & Saligrama, V.

Proceedings of ICML 2022, PMLR 162, 11694–11709.

Adaptive oracle-efficient online learning

Wang, G., Hu, Z., Muthukumar, V., & Abernethy, J. D.

Advances in Neural Information Processing Systems 35, 23398–23411.

Observation-Free Attacks on Stochastic Bandits

Xu, Y., Kumar, B., Abernethy, J.

In Advances in Neural Information Processing Systems 34 (NeurIPS 2021).

Understanding How Over-Parametrization Leads to Acceleration: A case of learning a single teacher neuron

Wang, J. K., Abernethy, J.

In Proceedings of the 13 th Asian Conference on Machine Learning. PMLR 157:17-32, 2021.

Fast convergence of fictitious play for diagonal payoff matrices

Abernethy, J. D., Lai, K., Wibisono, A.

In Proceedings of the 2021 ACM-SIAM Symposium on Discrete Algorithms (SODA).

A Modular Analysis of Provable Acceleration via Polyak’s Momentum: Training a Wide ReLU Network and a Deep Linear Network

Wang, J-K., Lin, C-H, Abernethy

In Proceedings of the 38th International Conference on Machine Learning, PMLR 139:10816-10827, 2021.

Last-iterate convergence rates for min-max optimization: convergence of Hamiltonian gradient descent and consensus optimization

Abernethy, J. D., Lai, K. A., & Wibisono, A.

Proceedings of ALT 2021, PMLR 132, 3–47.

Escaping saddle points faster with stochastic momentum

Wang, J.-K., Lin, C.-H., & Abernethy, J. D.

International Conference on Learning Representations (ICLR 2020).

Dynamic Online Pricing with Incomplete Information Using Multiarmed Bandit Experiments

Misra, K., Schwartz, E. M., & Abernethy, J.

Marketing Science.

Online learning via the differential privacy lens

Abernethy, J. D., Jung, Y. H., Lee, C., McMillan, A., & Tewari, A.

Advances in Neural Information Processing Systems 32, 8892–8902.

Learning auctions with robust incentive guarantees

Abernethy, J. D., Cummings, R., Kumar, B., Taggart, S., & Morgenstern, J.

Advances in Neural Information Processing Systems 32, 11587–11597.

Competing against Nash equilibria in adversarially changing zero-sum games

Rivera Cardoso, A., Abernethy, J. D., Wang, H., & Xu, H.

Proceedings of ICML 2019, PMLR 97, 921–930.

Acceleration through optimistic no-regret dynamics

Wang, J. K., & Abernethy, J. D.

In Advances in Neural Information Processing Systems (pp. 3824-3834).

ActiveRemediation: The Search for Lead Pipes in Flint, Michigan

Abernethy, J., Chojnacki, A., Farahi, A., Schwartz, E., & Webb, J.

In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 5-14). ACM.

Faster Rates for Convex-Concave Games

Abernethy, J., Lai, K., Levy, K. Y., & Wang, J. K.

In Proceedings of the 31st Conference On Learning Theory (Vol. 75, pp. 1595-1625). PMLR.

On frank-wolfe and equilibrium computation

Abernethy, J. D., & Wang, J. K.

In Advances in Neural Information Processing Systems (pp. 6584-6593).

A data science approach to understanding residential water contamination in flint

Chojnacki, A., Dai, C., Farahi, A., Shi, G., Webb, J., Zhang, D.T., Abernethy, J. and Schwartz, E.

In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1407-1416). ACM.

On convergence and stability of GANs

Kodali, N., Abernethy, J. D., Hays, J., & Kira, Z.

arXiv preprint arXiv:1705.07215.

Perturbation techniques in online learning and optimization

Abernethy, J. D., Lee, C., & Tewari, A.

In Perturbations, Optimization, and Statistics, 233–256.

Analysing Ratemyprofessors evaluations across institutions, disciplines, and cultures: The tell-tale signs of a good professor

Azab, M., Mihalcea, R., & Abernethy, J.

In International Conference on Social Informatics (pp. 438-453). Springer, Cham.

Utilizing high-dimensional features for real-time robotic applications: Reducing the curse of dimensionality for recursive bayesian estimation

Li, J., Ozog, P., Abernethy, J., Eustice, R. M., & Johnson-Roberson, M.

In 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 1230-1237). IEEE.

Threshold bandits, with and without censored feedback

Abernethy, J. D., Amin, K., & Zhu, R.

Advances in Neural Information Processing Systems 29, 4889–4897.

Rate of Price Discovery in Iterative Combinatorial Auctions

Abernethy, J., Lahaie, S., & Telgarsky, M.

In Proceedings of the 2016 ACM Conference on Economics and Computation (pp. 809-809). ACM.

Faster Convex Optimization: Simulated Annealing with an Efficient Universal Barrier

Abernethy, J., & Hazan, E.

In International Conference on Machine Learning (pp. 2520-2528).

Low-cost learning via active data procurement

Abernethy, J., Chen, Y., Ho, C., & Waggoner, B.

The Sixteenth ACM Conference on Economics and Computation (pp. 619-636). New York: ACM.

Financialized methods for market-based multi-sensor fusion

Abernethy, J., & Johnson-Roberson, M.

IEEE/RSJ International Conference on Intelligent Robots and Systems (pp. 900-907). Hamburg, Germany: IEEE.

Price Discovery in Subgradient Combinatorial Auctions

Abernethy, J., Lahaie, S., & Telgarsky, M.

The Third Conference on Auctions, Market Mechanisms and Their Applications. Chicago, Illinois.

A Market Framework for Eliciting Private Data

Waggoner, B., Frongillo, R., & Abernethy, J.

Advances in Neural Information Processing Systems 28 (NIPS 2015). Montreal, CA.

Fighting Bandits with a New Kind of Smoothness

Abernethy, J., Lee, C., Tewari, A.

Advances in Neural Information Processing Systems 28 (NIPS 2015). Montreal, CA.

On risk measures, market making, and exponential families

Abernethy, J. D., Frongillo, R. M., & Kutty, S.

ACM SIGecom Exchanges, 13(2), 21–25.

Jamming Defense Against a Resource-Replenishing Adversary in Multi-channel Wireless Systems

Wang, Q., Sheng, S., Abernethy, J., & Liu, M.

The 12th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks. Hammamet, Tunisia: IEEE.

Information aggregation in exponential family markets

Abernethy, J., Kutty, S., Lahaie, S., & Sami, R.

The Fifteenth ACM Conference on Economics and Computation (pp. 395-412). New York: ACM.

A general volume-parameterized market making framework

Abernethy, J., Frongillo, R., Li, X., Wotman Vaughan, J.

The Fifteenth ACM Conference on Economics and Computation (pp. 413-430). New York: ACM.

Online Linear Optimization via Smoothing

Abernethy, J., Lee, C., Sinha, A., & Tewari, A.

The 27 th annual Conference on Learning Theory June 13-15. Barcelona, Spain.

Efficient Market Making via Convex Optimization, and a Connection to Online Learning

Abernethy, J., Chen, Y., & Vaughan, J.

ACM Transactions on Economics and Computation, 1(2), 1-39.

How to Hedge an Option Against an Adversary: Black-Scholes Pricing is Minimax Optimal

Abernethy, J., Bartlett, P., Frongillo, R., & Wibisono, A.

Neural Information Processing Systems 2013 (pp. 2346-2354). Lake Tahoe, Nevada.

Adaptive Market Making via Online Learning

Abernethy, J., Kale, S.

Neural Information Processing Systems 2013 (pp. 2058-2066). Lake Tahoe, Nevada.

Minimax optimal algorithms for unconstrained linear optimization

McMahan, B., Abernethy, J.

Neural Information Processing Systems 2013 (pp. 2724-2732). Lake Tahoe, Nevada.

Large-scale bandit problems and KWIK learning

Abernethy, J. D., Amin, K., Draief, M., & Kearns, M.

Proceedings of the 30th International Conference on Machine Learning, PMLR 28, 588–596.

Interior-Point Methods for Full-Information and Bandit Online Learning

Abernethy, J., Hazan, E., & Rakhlin, A.

IEEE Transactions on Information Theory, 58(7), 4164-4175.

Minimax option pricing meets black-scholes in the limit

Abernethy, J., Frongillo, R., & Wibisono, A.

The 44th Symposium on Theory of Computing Conference (pp. 1029-1040). New York, New York: ACM.

A Characterization of Scoring Rules for Linear Properties

Abernethy, J., & Frongillo, R.

The 25 th Conference on Learning Theory (pp. 27.1-27.13). Edinburgh, Scotland.

A Collaborative Mechanism for Crowdsourcing Prediction Problems

Abernethy, J., & Frongillo, R.

Neural Information Processing Systems 2011 (pp. 2600-2608). Granada, Spain.

An optimization-based framework for automated market-making

Abernethy, J., Chen, Y., & Wortman Vaughan, J.

The 12th ACM Conference on Electronic Commerce (pp. 297-306). San Jose, California.

Blackwell approachability and no-regret learning are equivalent

Abernethy, J. D., Bartlett, P. L., & Hazan, E.

Proceedings of COLT 2011, PMLR 19, 27–46.

Does an efficient calibrated forecasting strategy exist? Proceedings of the 24th Annual Conference on Learning Theory (COLT), 809–812.

Abernethy, J. D., & Mannor, S.

Graph regularization methods for Web spam detection

Abernethy, J., Chapelle, O., & Castillo, C.

Machine Learning, 81(2), 207-225.

Repeated Games against Budgeted Adversaries

Abernethy, J., & Warmuth, M.

The 24th Annual Conference on Neural Information Processing Systems (pp. 1-9). Vancouver, British Columbia, Canada.

A regularization approach to metrical task systems

Abernethy, J., Bartlett, P., Buchbinder, N., & Stanton, I.

Algorithmic Learning Theory, 21st International Conference (pp. 270-284). Canberra, Australia.

Can we learn to gamble efficiently? Proceedings of the 23rd Annual Conference on Learning Theory (COLT), 318–319.

Abernethy, J. D.

A New Approach to Collaborative Filtering: Operator Estimation with Spectral Regularization

Abernethy, J., Bach, F., Evgeniou, T., & Vert, J.

Journal of Machine Learning Research, 10, 803-826.

A Stochastic View of Optimal Regret through Minimax Duality

Abernethy, J., Agarwal, A., Bartlett, P., & Rakhlin, A.

The 22nd Conference on Learning Theory. Montreal, Quebec, Canada.

Beating the adaptive bandit with high probability

Abernethy, J., Rakhlin, A.

The 22nd Conference on Learning Theory. Montreal, Quebec, Canada.

An efficient bandit algorithm for √T regret in online multiclass prediction

Abernethy, J. D., & Rakhlin, A.

Proceedings of the 22nd Annual Conference on Learning Theory (COLT).

Minimax games with bandits

Abernethy, J. D., & Warmuth, M. K.

Proceedings of the 22nd Annual Conference on Learning Theory (COLT).

Eliciting Consumer Preferences Using Robust Adaptive Choice Questionnaires

Abernethy, J., Evgeniou, T., Toubia, O., & Vert, J.

IEEE Transactions on Knowledge and Data Engineering, 20(2), 145-155.

When Random Play is Optimal Against an Adversary

Abernethy, J., Warmuth, M., & Yellin, J.

The 21st Annual Conference on Learning Theory (pp. 437-446). Helsinki, Finland.

Competing in the Dark: An Efficient Algorithm for Bandit Linear Optimization

Abernethy, J., Hazan, E., Rakhlin, A.

The 21st Annual Conference on Learning Theory (pp. 263-274). Helsinki, Finland.

Optimal Stragies and Minimax Lower Bounds for Online Convex Games

Abernethy, J., Bartlett, P., Rakhlin, A., Tewari, A.

The 21st Annual Conference on Learning Theory (pp. 415-424). Helsinki, Finland.

Web spam identification through content and hyperlinks

Abernethy, J., Chapelle, O., & Castillo, C.

The Fourth International Workshop on Adversarial Information Retrieval on the Web (pp. 41-44). New York, NY: ACM.

Online discovery of similarity mappings

Rakhlin, A., Abernethy, J., & Bartlett, P.

Machine Learning, Proceedings of the Twenty-Fourth International Conference (pp. 767-774). New York: ACM.

Multitask Learning with Expert Advice

Abernethy, J., Bartlett, P., & Rakhlin, A.

The 20th Annual Conference on Learning Theory (pp. 484-498). San Diego, CA.

Continuous experts and the binning algorithm

Abernethy, J. D., Langford, J., & Warmuth, M. K.

Proceedings of COLT 2006, 544–558. DOI: 10.1007/11776420_40.