Di He
助理教授
Researches machine learning, including Transformer and graph neural network architectures and ML for science such as molecular modeling.
Di He's lab at Peking works on graph learning and accelerators, chips & architecture. 30 papers since 2023, 22 at top venues, mostly ICLR and ICML. Applications close .
At a glance
- papers, 2023–2026
- 8/ year
- at T1 venues
- 73%
- PhD students
- —
- MS students
- —
- alumni listed
- —
Papers by yearfrom the lab homepage, 2023–2026
- 8 papers2026 · T1 5 · collecting
- 5 papers2025 · T1 3
- 9 papers2024 · T1 8
- 8 papers2023 · T1 6
T1 venuesOtherVenues: ICLR, ICML, AISTATS, NeurIPS, AAAI
Research focuspaper titles that use each topic's words
- Graph learning8 papers
- transformer2 papers
- Accelerators, chips & architecture1 paper
- network architectures0 papers
- neural network0 papers
Paper list
From the lab homepage.
20268 papers
- ICLR (T1)Diagnosing and Improving Diffusion Models by Estimating Optimal Loss Value (opens in a new tab)
- AAAI (T1)Efficient Reasoning for Large Reasoning Language Models via Certainty-Guided Reflection Suppression (opens in a new tab)
- EACLEnhancing Auto-regressive Chain-of-Thought through Loop-Aligned Reasoning (opens in a new tab)
- ICLR (T1)From Five Dimensions to Many: Large Language Models as Precise and Interpretable Psychological Profilers (opens in a new tab)
- ICLR Oral, top 1.18%In-Place Test-Time Training (opens in a new tab)
- ICLR (T1)Online Pseudo-Zeroth-Order Training of Neuromorphic Spiking Neural Networks (opens in a new tab)
- ICLR Oral, top 1.18%Quotient-Space Diffusion Model (opens in a new tab)
- ICLR (T1)Understanding vs. Generation: Navigating Optimization Dilemma in Multimodal Models (opens in a new tab)
20255 papers
- EMNLP (T1)Beyond Online Sampling: Bridging Offline-to-Online Alignment via Dynamic Data Transformation for LLMs (opens in a new tab)
- ICML Spotlight, top 2.59%DPO Meets PPO: Reinforced Token Optimization for RLHF (opens in a new tab)
- FPI-ICLRFollow Hamiltonian Leader: An Efficient Energy-Guided Sampling Method (opens in a new tab)
- ICLR (T1)Let the Code LLM Edit Itself When You Edit the Code (opens in a new tab)
- NeurIPS (T1)Theoretical Benefit and Limitation of Diffusion Language Model (opens in a new tab)
20249 papers
- ICLR Outstanding Paper Honorable MentionBeyond Weisfeiler-Lehman: A Quantitative Framework for GNN Expressiveness (opens in a new tab)
- NeurIPS (T1)Bridging Geometric States via Geometric Diffusion Bridge (opens in a new tab)
- ICML (T1)Do Efficient Transformers Really Save Computation? (opens in a new tab)
- ICML (T1)GeoMFormer: A General Architecture for Geometric Molecular Representation Learning (opens in a new tab)
- ICLR (T1)Hebbian Learning based Orthogonal Projection for Continual Learning of Spiking Neural Networks (opens in a new tab)
- AISTATS (T1)Learning a Fourier Transform for Linear Relative Positional Encodings in Transformers (opens in a new tab)
- NAACL (T1)REST: Retrieval-Based Speculative Decoding (opens in a new tab)
- ICML (T1)Temporal Spiking Neural Networks with Synaptic Delay for Graph Reasoning
- ICML (T1)Two Stones Hit One Bird: Bilevel Positional Encoding for Better Length Extrapolation (opens in a new tab)
20238 papers
- ICML (T1)A Complete Expressiveness Hierarchy for Subgraph GNNs via Subgraph Weisfeiler-Lehman Tests (opens in a new tab)
- AISTATS (T1)Adversarial Noises Are Linearly Separable for (Nearly) Random Neural Networks (opens in a new tab)
- ICLR (T1)Denoising Masked Autoencoders are Certifiable Robust Vision Learners (opens in a new tab)
- CVPR (T1)DSVT: Dynamic Sparse Voxel Transformer with Rotated Sets (opens in a new tab)
- AISTATS (T1)Learning Physics-Informed Neural Networks without Stacked Back-propagation (opens in a new tab)
- ICLR (T1)One Transformer Can Understand Both 2D & 3D Molecular Data (opens in a new tab)
- ICLR Outstanding Paper AwardRethinking the Expressive Power of GNNs via Graph Biconnectivity (opens in a new tab)
- NeurIPS Oral, top 0.7%Towards Revealing the Mystery behind Chain of Thought: a Theoretical Perspective (opens in a new tab)
Where alumni went
Alumni are not listed on the lab homepage.
Apply to PekingFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNPeking
Jun Sun
- Recommendation, search & web
- CNZhejiang
Jian Wu
- Recommendation, search & web
- CNZhejiang
Xiaolin Zheng
- Recommendation, search & web
- CNSJTU
Yanyan Shen
- Graph mining & networks
- CNTsinghua
Fei Qiao
- Accelerators, chips & architecture
- CAToronto
Ervin Sejdić
- Medical imaging & healthcare AI
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The member list could not be found on the lab site.
- The alumni list could not be found on the lab site.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Not collected
- Alumni
- Not collected
Tiers are venue classifications, not lab rankings.
Pekingcloses