Stanford Tabular and Relational (STAR) Project

Stanford Tabular and Relational (STAR) Project

Foundation models for tabular and relational data.

Advancing foundation models for structured data, from a single table to the many linked tables of a relational database.

Explore the research Models & datasets ↗

Latest

Jul 2026 · Conference

PluRel to appear at ICML 2026

The synthetic-data and scaling-laws work for relational foundation models is presented at the International Conference on Machine Learning.

Read more

Open Resources

Models, datasets, and code.

🤗 Hugging Face · stanford-starmodels & datasets GitHub · stanford-staropen-source code

Models

ModelDescriptionParams
rt-j Context-efficient relational foundation model pretrained on THE JOIN. 85M open ↗
rt-plurel Relational Transformer pretrained on the PluRel synthetic databases. 22M open ↗

Datasets

DatasetDescriptionDatabases
the-join The largest open relational corpus, with around 6,000 forecasting tasks. 650 open ↗
relbench Real-world relational databases with diverse predictive tasks. 7 open ↗
relbench-v2-extra Real-world databases extending the RelBench v2 collection. 3 open ↗
plurel Synthetic relational databases for scaling-law pretraining. 2,000 open ↗
redelex Databases ported from the CTU Prague Relational Learning Repository. 71 open ↗
tgb The Temporal Graph Benchmark of dynamic graphs for relational learning. 12 open ↗
dbinfer Databases from the 4DBInfer benchmark for graph-centric predictive modeling. 7 open ↗

Talks

Talks and presentations.

Talk: PluRel, Synthetic Data unlocks Scaling Laws for Relational Foundation Models PluRel: Synthetic Data unlocks Scaling Laws for Relational Foundation Models Vignesh Kothapalli · Temporal Graph Learning Reading Group Talk: Relational Transformer, Toward Zero-Shot Foundation Models for Relational Data Relational Transformer: Toward Zero-Shot Foundation Models for Relational Data Rishabh Ranjan · Temporal Graph Learning Reading Group Keynote by Jure Leskovec on relational foundation models at ACM RecSys Relational Foundation Models Jure Leskovec · ACM RecSys 2025 keynote Talk: Large Language Models are Good Relational Learners Large Language Models are Good Relational Learners Fang Wu & Vijay Prakash Dwivedi · Temporal Graph Learning Reading Group Talk: RelBench, A Benchmark for Deep Learning on Relational Databases RelBench: A Benchmark for Deep Learning on Relational Databases Rishabh Ranjan · Temporal Graph Learning Reading Group

People

Built at Stanford, with collaborators across academia and industry.

Faculty
JLJure Leskovec
Jure LeskovecProfessor of Computer Science · Stanford
CGCarlos Guestrin
Carlos GuestrinProfessor of Computer Science · Stanford
Core team
RRRishabh Ranjan
Rishabh RanjanPhD Student
VKVignesh Kothapalli
Vignesh KothapalliPhD Student
CKCharilaos Kanatsoulis
Charilaos KanatsoulisResearch Associate
Other contributors

Alejandro Dobles · Fang Wu · Federico López · Fengyu Li · Haiming Tang · Harshvardhan Agarwal · Jan Eric Lenssen · Jiaqi Han · Jiaxuan You · Joe Meyer · Johannes Hoffart · Joshua Robinson · Justin Gu · Kexin Huang · Mahmoud Mohammadi · Mark Znidar · Martin Jurkovic · Matthias Fey · Parth Shroff · Pranshu Chaturvedi · Rex Ying · Rishi Puri · Roshan Upendra · Shenyang Huang · Sri Jaladi · Tianlang Chen · Tom Palczewski · Valter Hudovernik · Vijay Prakash Dwivedi · Weihua Hu · Xinwei He · Yangyi Shen · Yiwen Yuan · Zecheng Zhang

Partner institutions
University of Oxford
Kumo AI
SAP
NVIDIA