Lei Xu
Zhiyuan Chair Professor
Works on statistical learning theory such as Bayesian Ying-Yang learning and its applications, with recent last-author work on self-supervised graph matching and alignment, target-aware molecule design and drug-target prediction.
Lei Xu's lab at SJTU works on learning theory & optimization, probabilistic & causal inference and interpretability, evaluation & safety of LMs. 48 papers since 2023, 20 at top venues, mostly AAAI and NeurIPS. 15 PhD students and 9 master's students. 16 alumni. Applications close .
At a glance
- papers, 2023–2026
- 13/ year
- at T1 venues
- 42%
- PhD students
- 15
- MS students
- 9
- alumni listed
- 16
Papers by yearfrom DBLP, 2023–2026
- 7 papers2026 · T1 5 · collecting
- 12 papers2025 · T1 5
- 15 papers2024 · T1 5
- 14 papers2023 · T1 5
T1 venuesOtherVenues: AAAI, NeurIPS, IJCAI, ICLR, KDD
Research focuspaper titles that use each topic's words
- Learning theory & optimization15 papers
- Probabilistic & causal inference5 papers
- molecule2 papers
- molecule design1 paper
- Interpretability, evaluation & safety of LMs0 papers
Paper list
From DBLP.
20267 papers
- Expert Systems with ApplicationsAn EEG-based dual-stream spatial-spectral-temporal large model for self-limited epilepsy with centrotemporal spikes (opens in a new tab)
- ICLR (T1)Beyond Instance-Level Alignment: Dual-Level Optimal Transport for Audio-Text Retrieval (opens in a new tab)
- AAAI (T1)Full-Atom Peptide Design via Riemannian-Euclidean Bayesian Flow Networks (opens in a new tab)
- TIP (T1)Harnessing Diffusion Models for Image Manipulation With Partial Sketches (opens in a new tab)
- PLoS Comput. Biol.PAIRNet: Predicting PIWI cleavage specificity via position-aware RNA interaction modeling (opens in a new tab)
- AAAI (T1)SEBSFormer: A Spectral-Enhanced Bi-Stream Transformer for Robust EEG Decoding (opens in a new tab)
- KDD (T1)UniRank: End-to-End Domain-Specific Reranking of Hybrid Text-Image Candidates (opens in a new tab)
202512 papers
- BIBMCATSyn: Predicting Synergistic Drug Combinations Through Context-Aware Heterogeneous Graph Convolution Model (opens in a new tab)
- Int. J. Approx. Reason.Cauchy-Schwarz bounded trade-off weighting for causal inference with small sample sizes (opens in a new tab)
- BIBMEnhancing piRNA Cleavage Prediction with piR-DANN: Integrating In Vitro and In Vivo Insights (opens in a new tab)
- J. Biomed. InformaticsGraphFusion: Integrative prediction of drug synergy using multi-scale graph representations and cell line contexts (opens in a new tab)
- NeurIPS (T1)KeeA*: Epistemic Exploratory A* Search via Knowledge Calibration (opens in a new tab)
- IEEE Trans. Comput. Biol. Bioinform.Multi-Objective Structure-Based Drug Design Using Causal Discovery (opens in a new tab)
- Neural NetworksPFB-Diff: Progressive Feature Blending diffusion for text-driven image editing (opens in a new tab)
- NeurIPS (T1)Prior-Guided Flow Matching for Target-Aware Molecule Design with Learnable Atom Number (opens in a new tab)
- KDD (T1)S3GA: Towards Scalable Self-Supervised Learning for Large Scale Graph Alignment (opens in a new tab)
- Machine LearningSketchMLP: effectively utilize rasterized images and drawing sequences for sketch recognition (opens in a new tab)
- NeurIPS (T1)Text to Sketch Generation with Multi-Styles (opens in a new tab)
- ICLR (T1)Towards Generalizable Reinforcement Learning via Causality-Guided Self-Adaptive Representations (opens in a new tab)
202415 papers
- IET Signal Process.A Deep Reinforcement Learning Approach for Portfolio Management in Non-Short-Selling Market (opens in a new tab)
- BIBMA Multimodal Deep Neural Network for Causally Learning Chest X-ray Images (opens in a new tab)
- Appl. Intell.Automated diagnosis of cervical spine physiological curvature based on deep neural networks with transformer by using nmODE (opens in a new tab)
- IJCAI (T1)Boosting Efficiency in Task-Agnostic Exploration through Causal Knowledge (opens in a new tab)
- IEEE ACM Trans. Comput. Biol. Bioinform.De Novo Drug Design by Multi-Objective Path Consistency Learning With Beam A* Search (opens in a new tab)
- Machine LearningIA-NGM: A bidirectional learning method for neural graph matching with feature fusion (opens in a new tab)
- Briefings Bioinform.KGDiff: towards explainable target-aware molecule generation with knowledge guidance (opens in a new tab)
- Comput. Vis. Image Underst.Lmser-pix2seq: Learning stable sketch representations for sketch healing (opens in a new tab)
- BIBMMedDet: Generative Adversarial Distillation for Efficient Cervical Disc Herniation Detection (opens in a new tab)
- AAAI (T1)Multilevel Attention Network with Semi-supervised Domain Adaptation for Drug-Target Prediction (opens in a new tab)
- BIBMNovelty Encouraged Beam Clustering Search for Multi-Objective De Novo Diverse Drug Design (opens in a new tab)
- NeurIPS (T1)SeeA*: Efficient Exploration-Enhanced A* Search by Selective Sampling (opens in a new tab)
202314 papers
- BIBMA Deep Learning Method with Multi-view Attention and Multi-branch GCN for BECT Diagnosis (opens in a new tab)
- ICASSP (T1)A Deep Temporal Factor Analysis Method for Large Scale Financial Portfolio Selection (opens in a new tab)
- PatternsA deep-learning method for the end-to-end prediction of intracranial aneurysm rupture risk (opens in a new tab)
- ICICA Dynamic Graph Convolutional Network for Anti-money Laundering (opens in a new tab)
- DATEDeepTH: Chip Placement with Deep Reinforcement Learning Using a Three-Head Policy Network (opens in a new tab)
- NeurIPS (T1)Generalized Weighted Path Consistency for Mastering Atari Games (opens in a new tab)
- ICONIPGeneralizing Graph Network Models for the Traveling Salesman Problem with Lin-Kernighan-Helsgaun Heuristics (opens in a new tab)
- IJCAI (T1)GLPocket: A Multi-Scale Representation Learning Approach for Protein Binding Site Prediction (opens in a new tab)
- Neural NetworksIA-FaceS: A bidirectional method for semantic face editing (opens in a new tab)
- AAAI (T1)Linking Sketch Patches by Learning Synonymous Proximity for Graphic Sketch Representation (opens in a new tab)
- BIBMMulti-source unsupervised domain-adaptation for automatic sleep staging (opens in a new tab)
- BIBMPEST: A General-Purpose Protein Embedding Model for Homology Search (opens in a new tab)
Where alumni went16 listed
Apply to SJTUFall 2027
Computer Science PhD (international applicants)
- Round
- Self-funded
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- SGNUS
Bryan Hooi
- Time series, anomaly & spatio-temporal data
- USUSC
Yan Liu
16 T1 - GBImperial
Deniz Gündüz
- Signal processing & information theory
- GBImperial
Francesco Belardinelli
- Reasoning & neuro-symbolic AI
- USStanford
James Zou
- Trustworthy & safe ML
- GBUCL
Siddharth Swaroop
- Probabilistic & causal inference
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- This site is shared with other labs, so counts may overlap.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
This lab shares a website with other labs. Member and alumni counts may include them.
Tiers are venue classifications, not lab rankings.
SJTUcloses