Niao He
Associate Professor
Leads the Optimization and Decision Intelligence group, studying large-scale stochastic and minimax optimization, reinforcement learning theory and probabilistic inference.
Niao He's lab at ETH Zurich works on learning theory & optimization, reinforcement learning and efficient training & inference. 68 papers since 2023, 56 at top venues, mostly NeurIPS and ICML. 11 PhD students. 12 alumni, most often to ByteDance and Jane Street.
At a glance
- papers, 2023–2026
- 18/ year
- at T1 venues
- 82%
- PhD students
- 11
- MS students
- 0
- alumni listed
- 12
Papers by yearfrom the lab homepage, 2023–2026
- 14 papers2026 · T1 11 · collecting
- 22 papers2025 · T1 17
- 19 papers2024 · T1 16
- 13 papers2023 · T1 12
T1 venuesOtherVenues: NeurIPS, ICML, AISTATS, ICLR, TMLR
Research focuspaper titles that use each topic's words
- Learning theory & optimization29 papers
- Reinforcement learning16 papers
- optimization15 papers
- Efficient training & inference9 papers
- probabilistic inference0 papers
Paper list
From the lab homepage.
202614 papers
- JMLR (T1)A Convex Framework for Confounding Robust Inference (opens in a new tab)
- Mathematical ProgrammingA Hessian-aware stochastic differential equation for modelling SGD
- ICLR (T1)A Schrodinger Eigenfunction Method for Long-Horizon Stochastic Optimal Control
- NeurIPS (T1)ANCRe: Adaptive Neural Connection Reassignment for Efficient Depth Scaling
- NeurIPS (T1)Global Optimality for Constrained Exploration via Penalty Regularization
- ICLR (T1)Landing with the Score: Riemannian Optimization through Denoising
- ICLR WorkshopManifold Generalization Provably Proceeds Memorization in Diffusion Models
- ICLR (T1)On the Benefits of Weight Normalization for Overparameterized Matrix Sensing
- OpticaOptical Computing with Spectrally Multiplexed Features in Complex Media
- ICML (T1)SALAAD: Sparse And Low-Rank Adaptation via ADMM for Large Language Model Inference
- NeurIPS (T1)Select-then-differentiate: Solving Bilevel Optimization with Manifold Lower-level Solution Sets
- NeurIPS (T1)Support Before Frequency in Discrete Diffusion
202522 papers
- ACM/IMS Journal of Data ScienceA Variational Inequality Approach to Independent Learning in Static Mean-Field Games (opens in a new tab)
- NeurIPS (T1)AmorLIP: Efficient Language-Image Pretraining via Amortization
- ICML (T1)Best of Both Worlds: Regret Minimization versus Minimax Play
- ICML (T1)Can RLHF be More Efficient with Imperfect Reward Models? A Policy Coverage Perspective
- JMLR (T1)EF21 with Bells & Whistles: Six Algorithmic Extensions of Modern Error Feedback (opens in a new tab)
- UAI (T1)Efficiently Escaping Saddle Points for Policy Optimization
- L4DCExploiting Approximate Symmetry for Efficient Multi-Agent Reinforcement Learning
- NeurIPS (T1)Flow Density Control: Generative Optimization Beyond Entropy-Regularized Fine-Tuning (opens in a new tab)
- AISTATS (T1)From Gradient Clipping to Normalization for Heavy Tailed SGD
- ICLR (T1)Learning to Steer Markovian Agents under Model Uncertainty
- SIAM Journal on Control and OptimizationLearning Zero-Sum Linear Quadratic Games with Improved Sample Complexity and Last-Iterate Convergence (opens in a new tab)
- NeurIPS (T1)Natural Gradient VI in Non-Conjugate Models
202419 papers
- NeurIPS (T1)Achieving Near-Optimal Convergence for Distributed Minimax Optimization with Adaptive Stepsizes (opens in a new tab)
- AAAI (T1)Automated Design of Affine Maximizer Mechanisms in Dynamic Settings (opens in a new tab)
- SIAM Journal on OptimizationConvergence of Entropy-Regularized Natural Policy Gradient with Linear Function Approximation (opens in a new tab)
- ICML (T1)DPZero: Private Fine-Tuning of Language Models without Backpropagation (opens in a new tab)
- Operations Research (T1)Efficient Algorithms for a Class of Stochastic Hidden Convex Optimization and Its Applications in Network Revenue Management (opens in a new tab)
- TMLR (T1)Finite-Time Analysis of Entropy-Regularized Neural Natural Actor-Critic Algorithm (opens in a new tab)
- SIAM Journal on Mathematics of Data ScienceFinite-Time Analysis of Natural Actor-Critic for POMDPs (opens in a new tab)
- AISTATS (T1)Generalization Bounds of Nonconvex-(Strongly)-Concave Stochastic Minimax Optimization (opens in a new tab)
- NeurIPS (T1)Implicit Regularization of Sharpness-Aware Minimization for Scale-Invariant Problems (opens in a new tab)
- AISTATS (T1)Independent Learning in Constrained Markov Potential Games (opens in a new tab)
- CDCIndependent Policy Mirror Descent for Markov Potential Games: Scaling to Large Number of Players (opens in a new tab)
- ICML (T1)Model-Based RL for Mean-Field Games is not Statistically Harder than Single-agent RL (opens in a new tab)
202313 papers
- SIAM Journal on Control and OptimizationA Discrete-time Switching System Analysis of Q-learning (opens in a new tab)
- AISTATS (T1)Kernel Conditional Moment Constraints for Confounding Robust Inference
- AISTATS (T1)Learning to Optimize for Stochastic Dominance Constraints (opens in a new tab)
- NeurIPS (T1)On Imitation in Mean-field Games (opens in a new tab)
- NeurIPS (T1)Optimal Guarantees for Algorithmic Reproducibility and Gradient Complexity in Convex Optimization (opens in a new tab)
- ICML (T1)Policy mirror ascent for efficient and independent learning in mean field games (opens in a new tab)
- TMLR (T1)Provably Convergent Policy Optimization via Metric-aware Trust Region Methods (opens in a new tab)
- ICML (T1)Reinforcement Learning with General Utilities: Simpler Variance Reduction and Large State-Action Space (opens in a new tab)
- NeurIPS (T1)Robust Knowledge Transfer in Tiered Reinforcement Learning (opens in a new tab)
- TAC (T1)Sample Complexity and Overparameterization Bounds for Temporal Difference Learning with Neural Network Approximation (opens in a new tab)
- ICML (T1)Stochastic Policy Gradient Methods: Improved Sample Complexity for Fisher-non-degenerate Policies (opens in a new tab)
- ICLR (T1)TiAda: A Time-scale Adaptive Algorithm for Nonconvex Minimax Optimization (opens in a new tab)
Where alumni went12 listed, 55% to industry
Industry 6Academia 5
Most often to
- ByteDance
- Jane Street
- Meta
Also to
- EPFL
- CUHK
- Nanjing University
Apply to ETH ZurichRolling
Not announced yet
Computer Science Doctorate (D-INFK)
- Round
- Rolling (no deadline)
- Opens
- —
- GRE
- See program page
- English
- Proficiency in written and spoken English
- Fee
- See program page
Confirmed on the program page.
Similar labsnearest by research summary
- CNTsinghua
Yu Liu
- Reinforcement learning
- USNYU
Qi Lei
- Learning theory & optimization
- USMichigan
Michał Dereziński
- Learning theory & optimization
- SGNTU
Lihui Chen
- Recognition, detection & segmentation
- USMIT
Tomaso Poggio
- Computational neuroscience & cognition
- CNTsinghua
Jian 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.