Profiles2020-2026
Mingyi Hong
University of Minnesota
Also published as M Hong
34Papers
15Last author
77Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers34
- Muon Outperforms Adam in Tail-End Associative Memory LearningAuthor 8 of 9ICLR 2026 · 40 citations
- StitchCUDA: An Automated Multi-Agents End-to-End GPU Programing Framework with Rubric-based Agentic Reinforcement LearningAuthor 5 of 6ICML 2026 · 9 citations
- HiPER: Hierarchical Plan–Execute RL for Multi-Turn LLM AgentsAuthor 7 of 7ICML 2026
- Scaling Unverifiable Rewards: A Case Study on Visual InsightsAuthor 5 of 7ACL Findings 2026
- DiSK: Differentially Private Optimizer with Simplified Kalman Filter for Noise ReductionAuthor 4 of 6ICLR 2025
- Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMsAuthor 2 of 5ICLR 2025
- Joint Reward and Policy Learning with Demonstrations and Human Feedback Improves AlignmentAuthor 8 of 8ICLR 2025
- On the Vulnerability of Applying Retrieval-Augmented Generation within Knowledge-Intensive Application DomainsAuthor 8 of 9ICML 2025
- RoSTE: An Efficient Quantization-Aware Supervised Fine-Tuning Approach for Large Language ModelsAuthor 7 of 7ICML 2025
- SLTrain: a sparse plus low rank approach for parameter and memory efficient pretrainingAuthor 4 of 8NeurIPS 2024 · 54 citations
- Getting More Juice Out of the SFT Data: Reward Learning from Human Demonstration Improves SFT for LLM AlignmentAuthor 6 of 6NeurIPS 2024 · 44 citations
- RAW: A Robust and Agile Plug-and-Play Watermark Framework for AI-Generated Images with Provable GuaranteesAuthor 4 of 5NeurIPS 2024 · 17 citations
- Differentially Private SGD Without Clipping Bias: An Error-Feedback ApproachAuthor 4 of 4ICLR 2024 · 15 citations
- Unraveling the Gradient Descent Dynamics of TransformersAuthor 5 of 5NeurIPS 2024 · 13 citations
- Demystifying Poisoning Backdoor Attacks from a Statistical PerspectiveAuthor 6 of 7ICLR 2024 · 11 citations
- MADA: Meta-Adaptive Optimizers Through Hyper-Gradient DescentAuthor 4 of 7ICML 2024 · 7 citations
- When Demonstrations meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement LearningAuthor 4 of 4NeurIPS 2023 · 33 citations
- A Unified Detection Framework for Inference-Stage Backdoor DefensesAuthor 6 of 7NeurIPS 2023 · 18 citations
- VCC: Scaling Transformers to 128K Tokens or More by Prioritizing Important TokensAuthor 3 of 7NeurIPS 2023 · 11 citations
- Understanding Clipping for Federated Learning: Convergence and Client-Level Differential PrivacyAuthor 3 of 5ICML 2022 · 134 citations
- Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time GuaranteesAuthor 4 of 4NeurIPS 2022 · 60 citations
- A Stochastic Linearized Augmented Lagrangian Method for Decentralized Bilevel OptimizationAuthor 8 of 8NeurIPS 2022 · 29 citations
- Inducing Equilibria via Incentives: Simultaneous Design-and-Play Ensures Global ConvergenceAuthor 5 of 7NeurIPS 2022 · 29 citations
- Distributed Optimization for Overparameterized Problems: Achieving Optimal Dimension Independent Communication ComplexityAuthor 5 of 5NeurIPS 2022 · 4 citations
- A Stochastic Multi-Rate Control Framework For Modeling Distributed Optimization AlgorithmsAuthor 2 of 4ICML 2022 · 1 citation
- A Near-Optimal Algorithm for Stochastic Bilevel Optimization via Double-MomentumAuthor 3 of 6NeurIPS 2021 · 175 citations
- STEM: A Stochastic Two-Sided Momentum Algorithm Achieving Near-Optimal Sample and Communication Complexities for Federated LearningAuthor 4 of 7NeurIPS 2021 · 78 citations
- Decentralized Riemannian Gradient Descent on the Stiefel ManifoldAuthor 3 of 4ICML 2021 · 64 citations
- When Expressivity Meets Trainability: Fewer than Neurons Can WorkAuthor 3 of 5NeurIPS 2021 · 11 citations
- Understanding Gradient Clipping in Private SGD: A Geometric PerspectiveAuthor 4 of 4NeurIPS 2020 · 254 citations
Showing the 30 most recent indexed papers.
