Haoxuan Li
助理教授
Researches machine learning, causal inference, trustworthy AI and large-model training and inference.
Haoxuan Li's lab at Peking works on probabilistic & causal inference, trustworthy & safe ML and efficient training & inference. 123 papers since 2023, 88 at top venues, mostly NeurIPS and ICML. Applications close .
At a glance
- papers, 2023–2026
- 32/ year
- at T1 venues
- 72%
- PhD students
- —
- MS students
- —
- alumni listed
- —
Papers by yearfrom the lab homepage, 2023–2026
- 52 papers2026 · T1 36 · collecting
- 41 papers2025 · T1 32
- 17 papers2024 · T1 12
- 13 papers2023 · T1 8
T1 venuesOtherVenues: NeurIPS, ICML, ICLR, KDD, SIGIR
Research focuspaper titles that use each topic's words
- Probabilistic & causal inference21 papers
- Efficient training & inference7 papers
- Trustworthy & safe ML4 papers
- inference3 papers
- large-model training0 papers
Paper list
From the lab homepage.
202652 papers
- ICLR (T1)A Relative Error-Based Evaluation Framework of Heterogeneous Treatment Effect Estimators (opens in a new tab)
- RecSys (Industry Track, Accepted Rate: 18.5%)A Self-Triggered Agentic Push Recommendation System
- TOISBatch-Adaptive Doubly Robust Learning for Debiasing Post-Click Conversion Rate Prediction Under Sparse Data (opens in a new tab)
- NeurIPS (T1)Bayesian Causal Experimental Design for CATE Estimation under Noncompliance
- ACMMMBeyond Isolated Clips: Ambiguity-Aware Global Memory for Video Social Relationship Recognition
- ICML (T1)Beyond Rational Illusion: Behaviorally Realistic Strategic Classification (opens in a new tab)
- TOISBeyond Similarity: Personalized Federated Recommendation with Composite Aggregation (opens in a new tab)
- ICLR (T1)Causal Discovery in the Wild: A Voting-Theoretic Ensemble Approach (opens in a new tab)
- ICML (T1)Causal Fine-Tuning under Latent Confounded Shift (opens in a new tab)
- ACM Computing SurveysCausal Inference with Complex Treatments: A Survey (opens in a new tab)
- TPAMI (T1)Causal-Theoretic Reward Modeling for RLHF from Observational Feedbacks
- KDD (T1)Causality-Based Conformal Imputation Correction with Non-Random Missing Labels (opens in a new tab)
202541 papers
- KDD (T1)A Two-Stage Pretraining-Finetuning Framework for Treatment Effect Estimation with Unmeasured Confounding (opens in a new tab)
- ICML (T1)Active Treatment Effect Estimation via Limited Samples (opens in a new tab)
- SIGIR (T1)Adaptive Structure Learning with Partial Parameter Sharing for Post-Click Conversion Rate Prediction (opens in a new tab)
- NeurIPS (T1)Breaking the Gradient Barrier: Unveiling Large Language Models for Strategic Classification (opens in a new tab)
- WWW (T1)CAP: Causal Air Quality Index Prediction Under Interference with Unmeasured Confounding (opens in a new tab)
- NeurIPS (T1)Causal Sufficiency and Necessity Improves Chain-of-Thought Reasoning (opens in a new tab)
- NAACL (T1)CharacterBox: Evaluating the Role-Playing Capabilities of LLMs in Text-Based Virtual Worlds (opens in a new tab)
- KDD (Full, Accepted Rate: 19%) #Classifying Treatment Responders: Bounds and Algorithms (opens in a new tab)
- NeurIPS (T1)Counterfactual Implicit Feedback Modeling (opens in a new tab)
- NeurIPS (T1)Curious Causality-Seeking Agents Learn Meta Causal World (opens in a new tab)
- TOISDebiased Recommendation via Wasserstein Causal Balancing (opens in a new tab)
- NeurIPS (T1)Decentralized Dynamic Cooperation of Personalized Models for Federated Continual Learning (opens in a new tab)
202417 papers
- ICML (T1)A Generative Approach for Treatment Effect Estimation under Collider Bias: From an Out-of-Distribution Perspective (opens in a new tab)
- NeurIPS (T1)A Local Method for Satisfying Interventional Fairness with Partially Known Causal Graphs (opens in a new tab)
- NeurIPS (T1)Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (opens in a new tab)
- ICLR (T1)Be Aware of the Neighborhood Effect: Modeling Selection Bias under Interference (opens in a new tab)
- AAAI (T1)Contrastive Balancing Representation Learning for Heterogeneous Dose-Response Curves Estimation (opens in a new tab)
- ICLR (Full, Accepted Rate: 5%)Debiased Collaborative Filtering with Kernel-Based Causal Balancing (opens in a new tab)
- KDD (Full, Accepted Rate: 20%)Debiased Recommendation with Noisy Feedback (opens in a new tab)
- ICML (Full, Accepted Rate: 3.6%)Learning Causal Relations from Subsampled Time Series with Two Time-Slices (opens in a new tab)
- ICML (T1)Learning Shadow Variable Representation for Treatment Effect Estimation under Collider Bias (opens in a new tab)
- ICLR (T1)MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlation (opens in a new tab)
- NeurIPS (T1)Neural Collapse Inspired Feature Alignment for Out-of-Distribution Generalization (opens in a new tab)
- ICML (T1)On the Maximal Local Disparity of Fairness-Aware Classifiers (opens in a new tab)
202313 papers
- RecSys (T1)ADRNet: A Generalized Collaborative Filtering Framework Combining Clinical and Non-Clinical Data for Adverse Drug Reaction Prediction (opens in a new tab)
- WWW (T1)Balancing Unobserved Confounding with a Few Unbiased Ratings in Debiased Recommendations (opens in a new tab)
- ICDM (Short, Accpeted Rate: 19.9%)CounterCLR: Counterfactual Contrastive Learning with Non-Random Missing Data in Recommendation (opens in a new tab)
- NeurIPS (Full, Accepted Rate: 26.1%) #Fairly Recommending with Social Attributes: A Flexible and Controllable Optimization Approach (opens in a new tab)
- AAAI (Full, Accepted Rate: 10.8%) #Multiple Robust Learning for Recommendation (opens in a new tab)
- NeurIPS (T1)Optimal Transport for Treatment Effect Estimation (opens in a new tab)
- ACMMM (Full, Accept Rate: 29.3%) #Pareto Invariant Representation Learning for Multimedia Recommendation (opens in a new tab)
- ICML (T1)Propensity Matters: Measuring and Enhancing Balancing for Recommendation (opens in a new tab)
- NeurIPS (T1)Removing Hidden Confounding in Recommendation: A Unified Multi-Task Learning Approach (opens in a new tab)
- ICLR (T1)StableDR: Stabilized Doubly Robust Learning for Recommendation on Data Missing Not at Random (opens in a new tab)
- ICLR (T1)TDR-CL: Targeted Doubly Robust Collaborative Learning for Debiased Recommendations (opens in a new tab)
- ICML (T1)Trustworthy Policy Learning under the Counterfactual No-Harm Criterion (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
- DETUM
Stephan Günnemann
- Trustworthy & safe ML
- CAToronto
Chris J. Maddison
- Probabilistic & causal inference
- CHEPFL
Ali H. Sayed
- Reasoning & neuro-symbolic AI
- SGNUS
Anji Liu
- Probabilistic & causal inference
- CNZhejiang
Xiaofei He
- Recommendation, search & web
- USMIT
Joshua Tenenbaum
- Computational neuroscience & cognition
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