Mehryar Mohri
Professor of Computer Science
Studies machine learning theory and algorithms, including learning guarantees, domain adaptation and weighted automata for speech and text.
Mehryar Mohri's lab at NYU works on learning theory & optimization. 46 papers since 2023, 1 at top venues, mostly EC. 2 PhD students. 22 alumni, most often to Google and Citadel. Applications close .
At a glance
- papers, 2023–2026
- 12/ year
- at T1 venues
- 2%
- PhD students
- 2
- MS students
- 0
- alumni listed
- 22
Papers by yearfrom the lab homepage, 2023–2026
- 15 papers2026 · T1 0 · collecting
- 9 papers2025 · T1 1
- 13 papers2024 · T1 0
- 9 papers2023 · T1 0
T1 venuesOtherVenues: EC
Research focuspaper titles that use each topic's words
- Learning theory & optimization17 papers
- guarantees6 papers
- adaptation3 papers
- domain adaptation2 papers
- domain2 papers
Paper list
From the lab homepage.
202615 papers
- Proceedings of the International Conference on Machine Learning (ICMLA theoretical framework for modular learning of robust generative models (opens in a new tab)
- Proceedings of The Annual Conference on Learning Theory (COLTAdaptive matrix online learning through smoothing with guarantees for nonsmooth nonconvex optimization (opens in a new tab)
- S802: ISAIMBalancing fairness: Learning with conflicting objectives and user feedback (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLBeyond binary: Continuous state optimization with graph-structured objectives (opens in a new tab)
- International Symposium on Artificial Intelligence and Mathematics (ISAIMBeyond Tsybakov: Model margin noise and H-consistency bounds (opens in a new tab)
- 95(1):5-21DC-programming for neural network optimizations (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLDistributional alignment games for answer-level fine-tuning (opens in a new tab)
- Proceedings of The International Conference on Algorithmic Learning Theory (ALTEfficient opportunistic approachabilityc (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLLinear-core surrogates: Smooth loss functions with linear rates for classification and structured prediction (opens in a new tab)
- 23(2):66-75Menus: A framework for learning against strategic opponents (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLMind the gap: Structure-aware consistency in preference learning (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLOptimized deferral for imbalanced settings (opens in a new tab)
20259 papers
- Proceedings of the International Conference on Machine Learning (ICMLBalancing the scales: A theoretical and algorithmic framework for learning from imbalanced data (opens in a new tab)
- Proceedings of The International Conference on Algorithmic Learning Theory (ALTEnhanced H-consistency bounds (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSHigh-dimensional calibration from swap regret (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSImproved balanced classification with theoretically grounded loss functions (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLMastering multiple-expert routing: Realizable H-consistency and strong guarantees for learning to defer (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLPrincipled algorithms for optimizing generalized metrics in binary classification (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSPrincipled model routing for unknown mixtures of source domains (opens in a new tab)
- Proceedings of The Annual Conference on Learning Theory (COLTRate-preserving reductions for Blackwell approachability (opens in a new tab)
- EC (T1)Swap regret and correlated equilibria beyond normal-form games (opens in a new tab)
202413 papers
- International Symposium on Artificial Intelligence and Mathematics (ISAIMA theory of learning with competing objectives and user feedback (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSA universal growth rate for learning with smooth surrogate losses (opens in a new tab)
- 92(2):393-438Best-effort adaptation (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLDifferentially private domain adaptation with theoretical guarantees (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLH-consistency guarantees for regression (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSMulti-label learning with stronger consistency guarantees (opens in a new tab)
- Proceedings of The International Conference on Algorithmic Learning Theory (ALTPredictor-rejector multi-class abstention: Theoretical analysis and algorithms (opens in a new tab)
- International Symposium on Artificial Intelligence and Mathematics (ISAIMPrincipled approaches for learning to defer with multiple experts (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSRealizable H-consistent and Bayes-consistent loss functions for learning to defer (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLRegression with multi-expert deferral (opens in a new tab)
- Twenty- Conference on Artificial Intelligence and Statistics (AISTATSTheoretically grounded loss functions and algorithms for score-based multi-class abstention (opens in a new tab)
- 92(2):277-315Theory and algorithms for learning with rejection in binary classification (opens in a new tab)
20239 papers
- Proceedings of the International Conference on Machine Learning (ICMLCross-entropy loss functions: Theoretical analysis and applications (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLH-consistency bounds for pairwise misranking loss surrogates (opens in a new tab)
- HNew Orleans, Louisiana, 2023
- Twenty- Conference on Artificial Intelligence and Statistics (AISTATSPrincipled approaches for private domain adaptation from a public source (opens in a new tab)
- Proceedings of the International Conference on Machine Learning (ICMLReinforcement learning can be more efficient with multiple rewards (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSStructured prediction with stronger consistency guarantees (opens in a new tab)
- Advances in Neural Information Processing Systems (NeurIPSTwo-stage learning to defer with multiple experts (opens in a new tab)
- Twenty- Conference on Artificial Intelligence and Statistics (AISTATSValencia, Spain, 2023
- Proceedings of The International Conference on Algorithmic Learning Theory (ALTvolume 201 of Proceedings of Machine Learning Research, pages
Where alumni went22 listed, 77% to industry
Industry 17Academia 5
Most often to
- Citadel
- Co-founder and Machine Learning Engineer at Unbox Research
Also to
- Courant Institute
- MIT
- UCLA
Apply to NYUFall 2027
Computer Science PhD (Courant / Tandon)
- Round
- Single round (December)
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- DETUM
Franziska Boenisch
- Trustworthy & safe ML
- USUCLA
Yuan Tian
- Trustworthy & safe ML
- USCMU
Maria-Florina Balcan
- Learning theory & optimization
- USUT Austin
Aryan Mokhtari
- Learning theory & optimization
- USUC Berkeley
Somayeh Sojoudi
- Trustworthy & safe ML
- CNSJTU
Xijun Li
- Learning theory & optimization
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
NYUcloses