Ofer Meshi's Homepage

I am a Research Scientist at Google. Previously I was a Research Assistant Professor at the Toyota Technological Institute at Chicago, a philanthropically endowed academic computer science institute at the University of Chicago.

I obtained my Ph.D. and M.Sc. in Computer Science from the Hebrew University of Jerusalem, where I worked with Amir Globerson and Nir Friedman. My B.Sc. in Computer Science is from Tel Aviv University.

My research is in Machine learning and optimization. In particular, I am interested in finding efficient algorithms for: recommendation systems, preference elicitation, reinforcement learning and other related problems.

Contact Information

Email:

Publications

• Asking Clarifying Questions for Preference Elicitation With Large Language Models
  A. Montazeralgahem, G. Tennenholz, C. Boutilier, O. Meshi
  In GENNEXT workshop at the Conference on Research and Development in Information Retrieval (SIGIR) 2025.

• Density-based User Representation using Gaussian Process Regression for Multi-interest Personalized Retrieval
  H. Wu, O. Meshi, M. Zoghi, F. Diaz, X. Liu, C. Boutilier, M. Karimzadehgan
  In Neural Information Processing Systems (NeurIPS) 2024.

• Model-Free Preference Elicitation
  C. Martin, C. Boutilier, T. Sandholm,  O. Meshi
  International Joint Conference on Artificial Intelligence (IJCAI) 2024

• Minimizing Live Experiments in Recommender Systems: User Simulation to Evaluate Preference Elicitation Policies
  C. Hsu, M. Mladenov, O. Meshi, J. Pine, H. Pham, S. Li, X. Liang, A. Polishko, L. Yang, B. Scheetz, C. Boutilier
  In Conference on Research and Development in Information Retrieval (SIGIR) 2024

• Model-Free Preference Elicitation
  C. Martin, C. Boutilier, O. Meshi
  In NeurIPS 2023 Workshop on Adaptive Experimental Design and Active Learning in the Real World.

• Preference Elicitation for Music Recommendations
  O. Meshi, J. Feldman, L. Yang, B. Scheetz, Y. Cai, M. Bateni, C. Salisbury, V. Aggarwal, C. Boutilier
  In ICML 2023 Workshop on Preference Learning.

• Overcoming Prior Misspecification in Online Learning to Rank
  J. Azizi, O. Meshi, M. Zoghi, M. Karimzadehgan
  In Artificial Intelligence and Statistics (AISTATS) 2023.

• On the Value of Prior in Online Learning to Rank
  B. Kveton, O. Meshi, Z. Qin, M. Zoghi
  In Artificial Intelligence and Statistics (AISTATS) 2022.

• Advantage Amplification in Slowly Evolving Latent-State Environments
  M. Mladenov, O. Meshi, J. Ooi, D. Schuurmans, C. Boutilier
  In International Joint Conference on Artificial Intelligence (IJCAI) 2019.

• Train and Test Tightness of LP Relaxations in Structured Prediction [external link]
  O. Meshi, B. London, A. Weller, D. Sontag
  Journal of Machine Learning Research (JMLR) 2019.

• MAP Estimation: Linear Programming Relaxation and Message-Passing Algorithms (Book chapter)
  O. Meshi and A. Schwing
  Handbook of Graphical Models, editors M. Maathuis, M. Drton, S. Lauritzen, M. Wainwright, CRC Press 2018.

• Seq2Slate: Re-ranking and Slate Optimization with RNNs
  I. Bello, S. Kulkarni, S. Jain, C. Boutilier, E. Chi, E. Eban, X. Luo, A. Mackey, O. Meshi
  ArXiv Preprint 2018.
  [short version] in ICML workshop on Negative Dependence in ML (2019)

• Deep Structured Prediction via Nonlinear Output Transforms
  C. Graber, O. Meshi, A. Schwing
  In Neural Information Processing Systems (NeurIPS) 2018.

• Planning and Learning with Stochastic Action Sets [Extended arXiv version]
  C. Boutilier, A. Cohen, A. Hassidim, Y. Mansour, O. Meshi, M. Mladenov, D. Schuurmans
  In International Joint Conference on Artificial Intelligence (IJCAI) 2018.

• Asynchronous Parallel Coordinate Minimization for MAP Inference
  O. Meshi and A. Schwing
  In Neural Information Processing Systems (NIPS) 2017.

• Logistic Markov Decision Processes
  M. Mladenov, C. Boutilier, D. Schuurmans, O. Meshi, G. Elidan, T. Lu
  In International Joint Conference on Artificial Intelligence (IJCAI) 2017.

• Linear-Memory and Decomposition-Invariant Linearly Convergent Conditional Gradient Algorithm for Structured Polytopes
  D. Garber and O. Meshi
  In Neural Information Processing Systems (NIPS) 2016.

• Bounding the Integrality Distance of LP Relaxations for Structured Prediction
  B. London, O. Meshi, A. Weller
  NIPS Workshop on Optimization for Machine Learning, 2016.

• Train and Test Tightness of LP Relaxations in Structured Prediction
  O. Meshi, M. Mahdavi, A. Weller, D. Sontag
  International Conference on Machine Learning (ICML) 2016.

• Fast and Scalable Structural SVM with Slack Rescaling
  H. Choi, O. Meshi, N. Srebro
  Artificial Intelligence and Statistics (AISTATS) 2016.

• On the Tightness of LP Relaxations for Structured Prediction
  O. Meshi, M. Mahdavi, D. Sontag
  NIPS Workshop on Optimization for Machine Learning, 2015.

• Smooth and Strong: MAP Inference with Linear Convergence
  O. Meshi, M. Mahdavi, A. Schwing
  Neural Information Processing Systems (NIPS) 2015.

• Efficient Training of Structured SVMs via Soft Constraints
  O. Meshi, N. Srebro, T. Hazan
  Artificial Intelligence and Statistics (AISTATS) 2015.

• Smoothed Coordinate Descent for MAP Inference (Book chapter)
  O. Meshi, T. Jaakkola, A. Globerson
  Advanced Structured Prediction, editors S. Nowozin, P. V. Gehler, J. Jancsary, C. Lampert, MIT Press 2014.

• Learning Structured Models with the AUC Loss and Its Generalizations  (supplementary)
  N. Rosenfeld, O. Meshi, D. Tarlow, A. Globerson
  Artificial Intelligence and Statistics (AISTATS) 2014.

• Learning Max-Margin Tree Predictors
  O. Meshi, E. Eban, G. Elidan, A. Globerson
  Uncertainty in Artificial Intelligence (UAI) 2013.

• Convergence Rate Analysis of MAP Coordinate Minimization Algorithms (supplementary)
  O. Meshi, T. Jaakkola, A. Globerson
  Neural Information Processing Systems (NIPS) 2012.

• An Alternating Direction Method for Dual MAP LP Relaxation
  O. Meshi and A. Globerson
  European Conference on Machine Learning (ECML PKDD) 2011.

• More data means less inference: A pseudo-max approach to structured learning (supplementary)
  D. Sontag, O. Meshi, T. Jaakkola, A. Globerson
  Neural Information Processing Systems (NIPS) 2010.

• Learning Efficiently with Approximate Inference via Dual Losses
  O. Meshi, D. Sontag, T. Jaakkola and A. Globerson
  International Conference on Machine Learning (ICML) 2010.

• FastInf: An Efficient Approximate Inference Library
  A. Jaimovich, O. Meshi, I. McGraw, G. Elidan
  Journal of Machine Learning Research (JMLR), 11:1733-1736, 2010.

• Convexifying the Bethe Free Energy
  O. Meshi, A. Jaimovich, A. Globerson and N. Friedman
  Uncertainty in Artificial Intelligence (UAI) 2009.

• Template Based Inference in Symmetric Relational Markov Random Fields
  A. Jaimovich, O. Meshi and N. Friedman
  Uncertainty in Artificial Intelligence (UAI) 2007.

• Evolutionary Conservation and over-representation of functionally enriched network patterns in the yeast regulatory network
  O. Meshi, T. Shlomi and E. Ruppin
  BMC Systems Biology, 1:1, 2007.

Theses

• Efficient Methods for Learning and Inference in Structured Output Prediction
  O. Meshi, PhD dissertation, 2013.

• Learning Symmetric Relational Markov Random Fields
  O. Meshi, MSc dissertation, 2007.

Personal

• Avital Meshi

• GitHub