Relational Deep Learning

Follow this topic Ask about this topic

Relational deep learning predicts labels or values for database entities by combining their attributes with information from linked rows and tables, which matters when outcomes depend on multi-hop context. Common approaches represent schemas as heterogeneous graphs and use message-passing GNNs, relational transformers, meta-path aggregation, or deep feature synthesis; row and temporal encoders capture table semantics and changing records. Relational foundation models add pretraining on real or synthetic databases and in-context prediction, commonly evaluated on multi-table classification and regression tasks such as RelBench. A central tension is whether richer neighborhoods and larger contexts justify their computational cost: flattening and simple feature propagation remain competitive, while sampling, schema choice, leakage control, and fair temporal evaluation shape reliable results.

Latest papers

49 papers8 papers in the last four weeks, level with the four weeks before. 0.1% of all new papers.

A Comparative Study of Importance Attribution to Columns in Deep Learning over Relational Databases
Nov 4, 2026Hila Levi-Yosefi, Shunit Agmon, Ester Livshits +2Feature Attribution Explainability Evaluation
Technion Haifa, Israel · Northeastern University Boston, USA
Speedrunning Relational Foundation Model Pretraining
Oct 20, 2026Gabriel Wedson Mendonça de Souza Veloso, Luan Borges Teodoro Reis Sena, Denis Oliveira Correa +1Relational Foundation Models Relational Deep Learning
Kunumi Institute, Belo Horizonte, Minas Gerais, Brazil · Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
Graph-to-Table Interfaces Control Relational Generalisation
Oct 19, 2026Alessandro Cagiano, Lorenzo Molfetta, Gianluca Moro +2Relational Deep Learning Tabular Prediction
University of Milan · University of Bologna · University of Edinburgh
Money-Mule Detection as a Relational Learning Problem: A Study on the RBIH NFPC Dataset
Oct 15, 2026Owais Ahmad, Anirudh DeodharFinancial Fraud Detection Anti-Money Laundering Detection
Applied Research, Quantiphi, Marlborough, MA 01752, USA
Revisiting Simple Baselines for Relational Deep Learning
Oct 14, 2026Viet The Nguyen, Vijay Prakash Dwivedi, Gunther GustGraph-Based Label Propagation Relational Deep Learning
University of Würzburg · Seamless Labs, Inc. · Stanford University
RT-J: Large-Scale Pretraining of Relational Transformers for Context-Efficient Predictions
Oct 13, 2026Rishabh Ranjan, Vignesh Kothapalli, Harshvardhan Agarwal +5Relational ICL Relational Foundation Models
Stanford University · SAP
Relational Learning as World Modeling
Oct 12, 2026Zheyu Zhang, Gjergji KasneciAction-Conditioned World Models World Model-Based Planning
Technical University of Munich Munich Center for Machine Learning (MCML)
DuRe-ST: Dual-Relation Spectro-Temporal Modeling for Speech Deepfake Detection
Oct 7, 2026Shaole Li, Siqing Qin, Youzhi Tu +1Audio Deepfake Detection ASVspoof
Department of Electrical and Electronic Engineering The Hong Kong Polytechnic University, Hong Kong SAR
ReGraph: A Computational Account of Emergent Generalization in the "what" and "where" Dual Visual Streams
Oct 6, 2026Hyewon Kang, Jungmin Lee, Ilgyu Lee +1Spatial Generalization Visual Neuroscience
Department of Intelligent Precision Healthcare Convergence, Sungkyunkwan University, Suwon, Republic of Korea · Center for Neuroscience Imaging Research, Institute for Basic Science (IBS), Suwon, Republic of Korea · Department of Biomedical Engineering, Sungkyunkwan University, Suwon, Republic of Korea
RelICL: Training-free Relational Learning with Tabular Foundation Models
Oct 1, 2026Simon Forbat, Rainer GemullaTabular Foundation Models Feature Interaction Modeling
Data and Web Science Group University of Mannheim Mannheim, Germany
STEER: Reducing Inference Cost in Relational Foundation Models through Semantically Informed Sampling
Oct 1, 2026Abdalla Mohamed, Ashraf AboulnagaRelational Foundation Models Efficient Inference
University of Texas at Arlington
Support-Set Target Leakage in Relational Foundation Models during In-Context Learning: Model Dependence and Evaluation Reliability
Sep 29, 2026Roshan Reddy Upendra, Alexandre Dorais, Joe Meyer +6Relational Foundation Models Data Leakage
SAP Palo Alto, CA, USA · SAP Montreal, Canada · SAP Seattle, WA, USA
Support-Set Target Leakage in Relational Foundation Models during In-Context Learning: Impact, Detection, and Mitigation
Sep 28, 2026Roshan Reddy Upendra, Alexandre Dorais, Joe Meyer +5Relational Foundation Models Data Leakage
SAP
Temporal Heterogeneous Graph Pretraining for Relational Deep Learning
Sep 28, 2026Yixin Peng, Er Jin, Diego Collarana +1Temporal GNNs Heterogeneous GNNs
RWTH Aachen University, Aachen, Germany · Fraunhofer FIT, Sankt Augustin, Germany
QUARTET: Quad-branch cross-Attention and Random-walk Traces for Enhancing Transformers on Relational Graphs
Sep 22, 2026Kyaw Hpone Myint, Nan Jiang, Xiang Li +4Temporal GNNs Transformer Attention
Capital One
MetaRTL: Meta-path Attention Enhanced Relational Table Learning
Sep 17, 2026Ken Zhong, Weichen Li, Zheng WangRelational Deep Learning Relational Databases
School of Computer Science, Shanghai Jiao Tong University, Shanghai, China · Shanghai Key Laboratory of Integrated Administration Technologies for Information Security, Shanghai, China
InRTL: Effective Intra-Inter Interaction Learning for Relational Tables
Sep 11, 2026Weichen Li, Ken Zhong, Zheng Wang +2Tabular Representation Learning Relational Deep Learning
Shanghai Jiao Tong University Shanghai, China
Relational Task Generation Language: A Declarative Specification Framework for Relational Deep Learning
Sep 1, 2026Oleksii Kolesnichenko, Jakub Peleška, Gustav ŠírRelational Deep Learning
Czech Technical University in Prague, Karlovo náměstí 13, Prague, 121 35, Czechia
Context Window Failures in Relational Foundation Models
Aug 31, 2026Denis Oliveira Correa, Francisco Galuppo AzevedoRelational Foundation Models Graph Foundation Models
Kunumi Institute, Brazil · Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil.
Can LLMs Use Relational Transformer Embeddings?
Aug 31, 2026Francisco Galuppo Azevedo, Clarissa Lima LouresTransformer LLM Fine-Tuning
Kunumi Institute, Brazil · Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil.
MetaSieve: Faster Relational Deep Learning through SQL-Based Metapath Selection
Aug 26, 2026Fahim Shahriar Khan, Ashraf AboulnagaRelational Deep Learning Graph Neural Networks
University of Texas at Arlington
RAD: Rule-Augmented Relational Anomaly Detection
Aug 24, 2026Noah Dahle, Anne Tumlin, Ngoc Tran +2Anomaly Detection Graph-Based Anomaly Detection
Vanderbilt University Nashville, Tennessee, USA
Advancing Open and Reproducible Relational Learning: RelArena- $α$ , TabPFN-Rel and RPI
Aug 17, 2026Adrian Hayler, Klemens Flöge, Alan Arazi +44Relational Deep Learning Relational Databases
Incremental Evaluation and Training in Relational Deep Learning
Aug 13, 2026Jakub Peleška, Gustav ŠírConcept Drift Online Continual Learning
Czech Technical University in Prague
EGRL: Edge generation-guided relation-aware learning for RNA-protein interaction prediction
Aug 13, 2026Danyu Li, Ling Zhou, Rubing Huang +3Graph Structure Learning Heterogeneous GNNs
School of Computer Science and Engineering, Macau University of Science and Technology, Macao SAR, 999078, China · Macau University of Science and Technology Zhuhai Research Institute, Zhuhai, Guangdong, 519099, China · Hubei Key Laboratory of Transportation Internet of Things, School of Computer Science and Artificial Intelligence, Wuhan University of Technology, Wuhan, Hubei, 430070, China +2
TRCoRSurg: Temporal-Relational Co-Reasoning for Surgical Video Triplet Recognition
Aug 5, 2026Fang Li, Shihao Zou, Weixin Si +3Surgical Video Understanding Relational Deep Learning
State Key Laboratory of Virtual Reality Technology and Systems, Beihang University · Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences · Faculty of Computer Science and Control Engineering, Shenzhen University of Advanced Technology
Curriculum Matters: Data-Efficient Relational PFN Pretraining with Synthetic Data
Jul 31, 2026Mohammad Sadeq Abolhasani, Viswanath GanapathySynthetic Data Pretraining Synthetic Data Generation
SAP Labs, LLC. Palo Alto, CA, United States
Structural Adversarial Attacks on Relational Deep Learning under Integrity Constraints
Jul 8, 2026Alan Gany, Bogdan Cautis, Silviu ManiuGraph Neural Networks Adversarial Attacks
Univ. Grenoble Alpes, CNRS, Grenoble INP, LIG Grenoble, France
A Fair Benchmarking of Deep Relational Database Learning Models
Jul 4, 2026Kazi F. Akhter, Bharath Ajendla, Manar D. SamadTransformer Graph Representation Learning
Department of Computer Science Tennessee State University Nashville, TN, USA · SAP AI Research and Innovation SAP Labs Palo Alto, CA, USA
Project Auto-World: Towards Automated Benchmarking of Neural Relational Reasoners
Jun 23, 2026Anirban Das, Joanne Boisson, Irtaza Khalid +2Language Model-Guided Search Heuristics Relational Reasoning
Cardiff University · University of Pennsylvania

Nov 4, 2026J/K move · Enter open · S save

A Comparative Study of Importance Attribution to Columns in Deep Learning over Relational Databases

Hila Levi-Yosefi, Shunit Agmon, Ester Livshits, Agapi Rissaki+1

Technion Haifa, Israel · Northeastern University Boston, USA

Feature Attribution Explainability Evaluation Shapley Value Attribution Relational Deep Learning

Open paper AskPDF arXiv

Deep learning over relational databases has led to increasingly effective prediction models, motivating a growing need to explain their behavior. We study explanation methods that assign importance scores to individual database attributes, focusing on a fundamental question: to what extent does the choice of importance-attribution method matter? We develop a methodology to systematically compare importance-attribution methods and analyze their similarities and differences. In particular, we study two aspects of the explanations: which attributes they deem important, captured by differences in attribute rankings, and how importance scores vary across attributes, captured by differences in score distributions. We further investigate whether these differences are reflected in the conventional measure of explanation quality, fidelity. We apply our methodology in an extensive empirical study of four methods: Monte Carlo Shapley, Kernel SHAP, single-attribute impact, and mask learning. We experiment with multiple datasets and tasks using two RDL architectures: heterogeneous GraphSAGE and Relational Graph Transformers. Our experiments show that explanation methods behave differently depending on the prediction type: regression or classification. In regression, most methods exhibit high agreement in attribute rankings, whereas classification yields considerably less consensus. Notably, while mask learning frequently diverges in its importance distributions, Shapley-based methods and single-attribute impact often converge to similar explanations. This suggests that, in many settings, single-attribute impact can replace Shapley at substantially lower computational cost. Overall, our study provides methodological tools for comparing importance-attribution methods, and empirical evidence that their choice has important consequences for interpreting relational deep learning models.