Research
My research develops graph-centric AI systems, spanning scalable graph learning, reasoning with large language models over structured knowledge, and applications to scientific and real-world domains. The directions below group my papers by topic. Full author lists and venues are on the publications page, and the homepage has an interactive map of the same topics.
Graph Learning
Graph machine learning methods that stay efficient, adaptive, and reliable on real-world graphs: scalable training and propagation for web-scale graphs, graph neural networks that adapt to heterogeneous structure, and analyses of bias, collapse, and generalization failures.
Scalability
Efficient optimization and propagation for learning on web-scale graphs.
- LazyGNN: Large-Scale Graph Neural Networks via Lazy Propagation ICML ’23 Code Authors and venue for LazyGNN
- Alternately Optimized Graph Neural Networks ICML ’23 Code Authors and venue for ALT-OPT
- Inverse APPNP: Solution for OGB-LSC 2022 MAG240M NeurIPS ’22 · Competition CodeResults Authors and venue for Inverse APPNP
Scalability papers on the publications page
Adaptivity
Models that adapt to heterogeneous nodes, links, and graph spectra.
- Mixture of Link Predictors on Graphs NeurIPS ’24 Code Authors and venue for Link-MoE
- Spectral-Aware Augmentation for Enhanced Graph Representation Learning CIKM ’24 Authors and venue for GASSER
- Demystifying Structural Disparity in Graph Neural Networks: Can One Size Fit All? NeurIPS ’23 Code Authors and venue for Structural Disparity
- Node-wise Filtering in Graph Neural Networks: A Mixture of Experts Approach Preprint · 2026 Authors and venue for Node-MoE
Adaptivity papers on the publications page
Reliability
Understanding and mitigating bias, collapse, and generalization failures in graph learning.
- Unveiling Mode Connectivity in Graph Neural Networks KDD ’25 Authors and venue for Mode Connectivity
- Structural Fairness-aware Active Learning for Graph Neural Networks ICLR ’24 Code Authors and venue for Structural Fairness
- Towards Label Position Bias in Graph Neural Networks NeurIPS ’23 Code Authors and venue for Label Position Bias
- Enhancing Graph Representations Learning with Decorrelated Propagation KDD ’23 Code Authors and venue for DeProp
- An Embarrassingly Simple Graph Heuristic Reveals Shortcut-Solvable Benchmarks for Sequential Recommendation NeurIPS ’26 · ED Primarily Recommendation & E-Commerce Code Authors and venue for Benchmark Audit (TGH)
Reliability papers on the publications page
Graphs + Foundation Models
How graphs and large language models strengthen each other: structured knowledge as the substrate for multi-step LLM reasoning and for retrieval-augmented generation (GraphRAG), and foundation models used as annotators, encoders, and predictors for graph data.
Structured LLM Reasoning
Explicit graph construction and exploration for reliable multi-step LLM reasoning.
- Reasoning by Exploration: A Unified Approach to Retrieval and Generation over Graphs WWW ’26 · Oral Code Authors and venue for Reasoning by Exploration
- Reasoning with Graphs: Structuring Implicit Knowledge to Enhance LLMs Reasoning ACL Findings ’25 Authors and venue for Reasoning with Graphs
- ConnectionMind: Leveraging Social Networks and Large Language Models for Personalized Recommendation at Meta CIKM ’26 · Applied Primarily Recommendation & E-Commerce Authors and venue for ConnectionMind
Structured LLM Reasoning papers on the publications page
GraphRAG
Retrieval, filtering, integration, evaluation, and security over structured knowledge.
- Beyond Static Retrieval: Opportunities and Pitfalls of Iterative Retrieval in GraphRAG EMNLP ’26 · Main Authors and venue for Beyond Static Retrieval
- Why Retrieval-Augmented Generation Fails: A Graph Perspective KDD ’26 Authors and venue for Why RAG Fails
- RAG vs. GraphRAG: A Systematic Evaluation and Key Insights KDD ’26 · D&B Code Authors and venue for RAG vs. GraphRAG
- Benchmarking Knowledge-Extraction Attack and Defense on Retrieval-Augmented Generation KDD ’26 · D&B · Oral CodeDataset Authors and venue for RAG Attack & Defense
- Empowering GraphRAG with Knowledge Filtering and Integration EMNLP ’25 Authors and venue for GraphRAG-FI
- Mixture of Structural-and-Textual Retrieval over Text-rich Graph Knowledge Bases ACL Findings ’25 Code Authors and venue for MoR
- Retrieval-Augmented Generation with Graphs (GraphRAG) Preprint · 2025 Resources Authors and venue for GraphRAG Survey
- From Flat to Structural: Enhancing Automated Short Answer Grading with GraphRAG EDM ’26 Primarily Education Authors and venue for GraphRAG for Grading
- Enhancing LLM-Based Short Answer Grading with Retrieval-Augmented Generation EDM ’25 Primarily Education Authors and venue for RAG for Grading
GraphRAG papers on the publications page
Foundation Models for Graphs
Language and video models as graph annotators, encoders, and predictors.
- Graph2Video: Leveraging Video Models to Model Dynamic Graph Evolution AAAI ’26 Authors and venue for Graph2Video
- GT2VEC: Large Language Models as Multi-Modal Encoders for Text and Graph-Structured Data KDD SKnow-LLM Workshop ’25 Code Authors and venue for GT2Vec
- Label-free Node Classification on Graphs with Large Language Models (LLMs) ICLR ’24 Code Authors and venue for LLM-GNN
- Understanding Graph Self-Supervised Pre-training under Distribution Shifts: A Scaling Law Perspective NeurIPS ’26 Authors and venue for Graph SSL Scaling
Foundation Models for Graphs papers on the publications page
Real-World Impact
Applications of graph-centric AI to recommendation and e-commerce, education, and science.
Recommendation & E-Commerce
Graph-based personalization, multimodal recommendation, and benchmark diagnostics.
- ConnectionMind: Leveraging Social Networks and Large Language Models for Personalized Recommendation at Meta CIKM ’26 · Applied Authors and venue for ConnectionMind
- Enhancing ID and Text Fusion via Alternative Training in Session-based Recommendation IJCNLP–AACL ’25 Code Authors and venue for AlterRec
- Amazon-M2: A Multilingual Multi-locale Shopping Session Dataset for Recommendation and Text Generation NeurIPS ’23 · D&B CodeAmazon ScienceDataset Authors and venue for Amazon-M2
- STGCN: A Spatial-Temporal Aware Graph Learning Method for POI Recommendation ICDM ’20 Authors and venue for STGCN
- An Embarrassingly Simple Graph Heuristic Reveals Shortcut-Solvable Benchmarks for Sequential Recommendation NeurIPS ’26 · ED Code Authors and venue for Benchmark Audit (TGH)
Recommendation & E-Commerce papers on the publications page
Education
Knowledge-grounded assessment design and automated grading.
- From Flat to Structural: Enhancing Automated Short Answer Grading with GraphRAG EDM ’26 Authors and venue for GraphRAG for Grading
- Human-AI Collaboration for Knowledge-in-Use Assessment Design: Leveraging LLMs with RAG IEEE TALE ’25 Authors and venue for Human–AI Assessment Design
- Enhancing LLM-Based Short Answer Grading with Retrieval-Augmented Generation EDM ’25 Authors and venue for RAG for Grading
- Constructing Educational Concept Maps with Multiple Relationships from Multi-source Data ICDM ’19 Authors and venue for Educational Concept Maps
Education papers on the publications page
AI for Science
Graph learning for multimodal single-cell data integration.
- Multimodal Single-Cell Data Integration Project · NeurIPS ’21 Competition · 1st CompetitionProject