Mingsheng Long
Associate Professor
Works on machine learning, especially transfer learning, deep learning, and knowledge learning for AI and industrial data software.
Mingsheng Long's lab at Tsinghua works on knowledge, QA & retrieval. 56 papers since 2023, 54 at top venues, mostly ICML and NeurIPS. 8 PhD students and 9 master's students. 33 alumni, most often to ByteDance and Alibaba. Applications close .
At a glance
- papers, 2023–2026
- 15/ year
- at T1 venues
- 96%
- PhD students
- 8
- MS students
- 9
- alumni listed
- 33
Papers by yearfrom the lab homepage, 2023–2026
- 10 papers2026 · T1 10 · collecting
- 13 papers2025 · T1 13
- 18 papers2024 · T1 18
- 15 papers2023 · T1 13
T1 venuesOtherVenues: ICML, NeurIPS, ICLR, TPAMI, JMLR
Research focuspaper titles that use each topic's words
- transfer4 papers
- transfer learning1 paper
- Knowledge, QA & retrieval0 papers
- software0 papers
- especially0 papers
Paper list
From the lab homepage.
202610 papers
- ICML (T1)Brep2Shape: Boundary and Shape Representation Alignment via Self-Supervised Transformers
- TPAMI (T1)Deep Time Series Models: A Comprehensive Survey and Benchmark
- ICLR (T1)Foresight Diffusion: Improving Sampling Consistency in Predictive Diffusion Models
- ICML (T1)GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training
- ICML (T1)It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarks
- ICLR (T1)Mitigating Noise Shift in Denoising Generative Models with Noise Awareness Guidance
- TMLR (T1)ProPINN: Demystifying Propagation Failures in Physics-Informed Neural Networks
- TKDE (T1)TimeAgent: A Training-Free LLM Agent for Time Series Foundation Models
- ICML (T1)Transolver-3: Scaling Up Transformer Solvers to Industrial-Scale Geometries
- ICLR (T1)Vid2World: Crafting Video Diffusion Models to Interactive World Models
202513 papers
- KDD (T1)CompilerDream: Learning a Compiler World Model for General Code Optimization
- JMLR (T1)depyf: Open the Opaque Box of PyTorch Compiler for Machine Learning Researchers
- ICLR (T1)Domain Guidance: A Simple Transfer Approach for a Pre-trained Diffusion Model
- ICLR (T1)Dynamical Diffusion: Learning Temporal Dynamics with Diffusion Models
- NeurIPS (T1)FlashBias: Fast Computation of Attention with Bias
- ICLR (T1)Long-Sequence Recommendation Models Need Decoupled Embeddings
- NeurIPS (T1)PhySense: Sensor Placement Optimization for Accurate Physics Sensing
- NeurIPS (T1)RLVR-World: Training World Models with Reinforcement Learning
- ICML (T1)Sundial: A Family of Highly Capable Time Series Foundation Models
- ICLR (T1)Timer-XL: Long-Context Transformers for Unified Time Series Forecasting
- ICML (T1)Trajectory World Models for Heterogeneous Environments
- ICML (T1)Transolver++: An Accurate Neural Solver for PDEs on Million-Scale Geometries
202418 papers
- NeurIPS (T1)AutoTimes: Autoregressive Time Series Forecasters via Large Language Models
- ICML (T1)CogDPM: Diffusion Probabilistic Models via Cognitive Predictive Coding
- NeurIPS (T1)DeepLag: Discovering Deep Lagrangian Dynamics for Intuitive Fluid Prediction
- NeurIPS (T1)Diffusion Tuning: Transferring Diffusion Models via Chain of Forgetting
- ICLR (T1)Efficient ConvBN Blocks for Transfer Learning and Beyond
- ICML (T1)HarmonyDream: Task Harmonization Inside World Models
- ICML (T1)HelmFluid: Learning Helmholtz Dynamics for Interpretable Fluid Prediction
- ICLR (T1)iTransformer: Inverted Transformers Are Effective for Time Series Forecasting
- NeurIPS (T1)iVideoGPT: Interactive VideoGPTs are Scalable World Models
- ICML (T1)Mobile Attention: Mobile-Friendly Linear-Attention for Vision Transformers
- ICML (T1)On the Embedding Collapse when Scaling up Recommendation Models
- TPAMI (T1)One Fits Many: Class Confusion Loss for Versatile Domain Adaptation
202315 papers
- ECMLBi-Tuning: Efficient Transfer from Pre-Trained Models
- ICML (T1)CLIPood: Generalizing CLIP to Out-of-Distributions
- ICML (T1)Estimating Heterogeneous Treatment Effects: Mutual Information Bounds and Learning Algorithms
- NeurIPS (T1)ForkMerge: Mitigating Negative Transfer in Auxiliary-Task Learning
- TPAMI (T1)From Big to Small: Adaptive Learning to Partial-Set Domains
- —IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 45(12):15275-15291, 2023 [Link] [Code]
- Nature Machine Intelligence (T1)Interpretable Weather Forecasting for Worldwide Stations with a Unified Deep Model
- NeurIPS (T1)Koopa: Learning Non-stationary Time Series Dynamics with Koopman Predictors
- TPAMI (T1)ModeRNN: Harnessing Spatiotemporal Mode Collapse in Unsupervised Predictive Learning
- NeurIPS (T1)Pre-training Contextualized World Models with In-the-wild Videos for Reinforcement Learning
- TPAMI (T1)PredRNN: A Recurrent Neural Network for Spatiotemporal Predictive Learning
- NeurIPS (T1)SimMTM: A Simple Pre-Training Framework for Masked Time-Series Modeling
Where alumni went33 listed, 60% to industry
Industry 12Academia 8
Most often to
- ByteDance
- Alibaba
- Kling AI
Also to
- CUHK
- East China Normal University
- Harbin Institute of Technology
Apply to TsinghuaFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- GRE
- See program page
- English
- Per school catalog
- Fee
- RMB 800
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNZhejiang
Zemin Liu
- Graph learning
- USCMU
Kun Zhang
- Probabilistic & causal inference
- HKCUHK
Qiang Xu
- Trustworthy & safe ML
- CNZhejiang
Lei Cheng
- Learning theory & optimization
- GBUCL
Anthony Bourached
- Signal processing & information theory
- USMIT
Tommi Jaakkola
- 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.
Tsinghuacloses