mlsanityData to 13 Oct 2026

Distribution Shift Aware Neural Tabular Learning

arXiv 2508.19486 ↗First posted 27 Aug 2025cs.LGcited by 4

Not on the map. Why?

Abstract (arXiv)

Tabular learning transforms raw features into optimized spaces for downstream tasks, but its effectiveness deteriorates under distribution shifts between training and testing data. We formalize this challenge as the Distribution Shift Tabular Learning (DSTL) problem and propose a novel Shift-Aware Feature Transformation (SAFT) framework to address it. SAFT reframes tabular learning from a discrete search task into a continuous representation-generation paradigm, enabling differentiable optimization over transformed feature sets. SAFT integrates three mechanisms to ensure robustness: (i) shift-resistant representation via embedding decorrelation and sample reweighting, (ii) flatness-aware generation through suboptimal embedding averaging, and (iii) normalization-based alignment between training and test distributions. Extensive experiments show that SAFT consistently outperforms prior tabular learning methods in terms of robustness, effectiveness, and generalization ability under diverse real-world distribution shifts.

Key facts. Cited by 4 papers in the map (data through 13 Oct 2026). Builds on 28 of its 45 references that are in the map. First posted on arXiv 27 Aug 2025.

Builds on

28 of 45 references.

References · 28

  1. 2017Inductive Representation Learning on Large Graphscited by 6,303
  2. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,202
  3. 2018Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNscited by 586
  4. 2014OpenML: networked science in machine learningcited by 524
  5. 2023Data-centric Artificial Intelligence: A Surveycited by 202
  6. 2017Feature Engineering for Predictive Modeling using Reinforcement Learningcited by 78
  7. 2019The autofeat Python Library for Automated Feature Engineering and Selectioncited by 73
  8. 2019Stable Learning via Sample Reweightingcited by 70
  9. 2023Continual Learning on Dynamic Graphs via Parameter Isolationcited by 36
  10. 2022Group-wise Reinforcement Feature Generation for Optimal and Explainable Representation Space Reconstructioncited by 35
  11. 2023Self-optimizing Feature Generation via Categorical Hashing Representation and Hierarchical Reinforcement Crossingcited by 30
  12. 2024Evolutionary Large Language Model for Automated Feature Transformationcited by 29
16 more
  1. 2023Distribution Shift Inversion for Out-of-Distribution Predictioncited by 23
  2. 2023Reinforcement-Enhanced Autoregressive Feature Transformation: Gradient-steered Search in Continuous Space for Postfix Expressionscited by 23
  3. 2024Feature Selection as Deep Sequential Generative Learningcited by 21
  4. 2024Knockoff-Guided Feature Selection via A Single Pre-trained Reinforced Agentcited by 15
  5. 2025CT-PatchTST: Channel-Time Patch Time-Series Transformer for Long-Term Renewable Energy Forecastingcited by 15
  6. 2024Neuro-Symbolic Embedding for Short and Effective Feature Selection via Autoregressive Generationcited by 14
  7. 2025A Survey on Data-Centric AI: Tabular Learning from Reinforcement Learning and Generative AI Perspectivecited by 14
  8. 2024Reinforcement Feature Transformation for Polymer Property Performance Predictioncited by 13
  9. 2025Sculpting Features from Noise: Reward-Guided Hierarchical Diffusion for Task-Optimal Feature Transformationcited by 13
  10. 2024Topology-aware Reinforcement Feature Space Reconstruction for Graph Datacited by 11
  11. 2025Unsupervised Feature Transformation via In-context Generation, Generator-critic LLM Agents, and Duet-play Teamingcited by 11
  12. 2023AdaER: An Adaptive Experience Replay Approach for Continual Lifelong Learningcited by 9
  13. 2025Agentic Feature Augmentation: Unifying Selection and Generation with Teaming, Planning, and Memoriescited by 9
  14. 2025Enhancing Customer Contact Efficiency with Graph Neural Networks in Credit Card Fraud Detection Workflowcited by 8
  15. 2025LLM-ML Teaming: Integrated Symbolic Decoding and Gradient Search for Valid and Stable Generative Feature Transformationcited by 5
  16. 2022Decorrelate Irrelevant, Purify Relevant: Overcome Textual Spurious Correlations from a Feature Perspectivecited by 4

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2025: 3; 2026: 1.

4 papers.

  1. 2025Permutation-Invariant Representation Learning for Robust and Privacy-Preserving Feature Selectioncited by 4
  2. 2025DELTA: Variational Disentangled Learning for Privacy-Preserving Data Reprogrammingcited by 3
  3. 2025Dataforge: Agentic Platform for Autonomous Data Engineeringcited by 1
  4. 2026Evolving Demonstration Optimization for Chain-of-Thought Feature Transformationcited by 0