Synthetic Data Pretraining

Ask about this topic

Overview

The field studies how models can learn useful representations, behaviors, or task-solving strategies from generated examples before being adapted to real tasks. Common approaches include physics-based simulation and procedural generation, synthetic text or code derived from source material, and prior-fitted models trained on generated tasks such as tables, graphs, and time series. Results show that synthetic pretraining can improve data efficiency and transfer, especially in specialized or low-data settings, but gains depend on the generator, curriculum, and match between training tasks and deployment. A central tension is how to scale synthetic data without teaching misleading patterns, losing diversity, or mistaking strong synthetic-benchmark performance for real-world generalization; many settings still benefit from real-data fine-tuning or careful verification.

Latest papers

CardsList

122 papers27 papers in the last four weeks, up 50% on the four weeks before. 0.2% of all new papers.

May 7, 2026

Mind the Gap? A Distributional Comparison of Real and Synthetic Priors for Tabular Foundation Models

Tabular foundation models are pre-trained on one of three classes of corpus: curated datasets drawn from benchmark repositories, tables harvested at scale from the web, or synthetic tables sampled from a parametric generative prior. Despite the centrality of pre-training data to model performance, little is known about how these corpora relate to one another in distribution, and the impact this has on downstream performance. In this work we take three canonical, archetypal datasets used to train tabular foundation models; the T4 dataset represents web-scraped corpora, the TabFM dataset curated tables from Kaggle, and the TabICL dataset as the only well-used synthetic prior with publicly available parameters. We characterise each corpus using aggregate features over whole tables, columns and correlations, and compare them using discriminator AUCs and k-NN coverage metrics. We find that the TabICL synthetic prior occupies a narrow region of the space of real tables, that this mismatch cannot be closed by optimising prior hyper-parameters across more than 86 thousand configurations, and that curated and web-scraped corpora are broadly interchangeable on a distributional level in feature space. Surprisingly, the distributional gap between synthetic pre-training data and real tables has a clearly detectable effect on performance under neither feature-based proximity measures or TabICL's own internal representations, suggesting that coverage of the real-data distribution is not the primary driver of TabICL's generalisation.

Alex O. Davies, Telmo de Menezes e Silva Filho, Nirav Ajmeri

School of Geographical Sciences University of Bristol, UK · School of Engineering Mathematics and Technology University of Bristol, UK · School of Computer Science University of Bristol, UK

Synthetic Data PretrainingSynthetic-to-Real Domain AdaptationTabular Foundation ModelsSynthetic Data Evaluation
AskPDF
May 5, 2026

Distribution-Free Pretraining of Classification Losses via Evolutionary Dynamics

We propose Evolutionary Dynamic Loss (EDL), a framework that learns a transferable classification loss in the probability space using unlimited synthetic prediction-label pairs, without accessing real samples during the main loss pretraining stage. EDL parameterizes the loss as a lightweight network and is trained with a semantics-free ranking-consistency objective that assigns larger penalties for more erroneous predictions. To robustly explore the space of loss functions, we optimize EDL via an evolutionary strategy and introduce chaotic mutation to improve exploration under noisy fitness evaluations. Experiments on CIFAR-10 with ResNet backbones show that EDL can serve as a drop-in replacement for cross-entropy and achieves competitive or improved accuracy, while ablation studies confirm that chaotic mutation yields faster convergence and better synthetic pretraining metrics than standard Gaussian mutation.

Meng Xiang, Yan Pei

School of Computer Science and Engineering University of Aizu Aizuwakamatsu, Japan

Synthetic Data PretrainingImage Classification
AskPDF
Apr 21, 2026

Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors

One of the most challenging problems in graph machine learning is generalizing across graphs with diverse properties. Graph neural networks (GNNs) face a fundamental limitation: they require separate training for each new graph, preventing universal generalization across diverse graph datasets. A critical challenge facing GNNs lies in their reliance on labeled training data for each individual graph, a requirement that hinders the capacity for universal node classification due to the heterogeneity inherent in graphs -- differences in homophily levels, community structures, and feature distributions across datasets. Inspired by the success of large language models (LLMs) that achieve in-context learning through massive-scale pre-training on diverse datasets, we introduce NodePFN. This universal node classification method generalizes to arbitrary graphs without graph-specific training. NodePFN learns posterior predictive distributions (PPDs) by training only on thousands of synthetic graphs generated from carefully designed priors. Our synthetic graph generation covers real-world graphs through the use of random networks with controllable homophily levels and structural causal models for complex feature-label relationships. We develop a dual-branch architecture combining context-query attention mechanisms with local message passing to enable graph-aware in-context learning. Extensive evaluation on 23 benchmarks demonstrates that a single pre-trained NodePFN achieves 71.27 average accuracy. These results validate that universal graph learning patterns can be effectively learned from synthetic priors, establishing a new paradigm for generalization in node classification.

Jeongwhan Choi, Jongwoo Kim, Woosung Kang +1

KAIST Daejeon, Republic of Korea

Synthetic Data PretrainingOOD GeneralizationIn-Context LearningGraph Foundation ModelsNode Classification
AskPDF
Apr 15, 2026

How Can We Synthesize High-Quality Pretraining Data? A Systematic Study of Prompt Design, Generator Model, and Source Data

Synthetic data is a standard component in training large language models, yet systematic comparisons across design dimensions, including rephrasing strategy, generator model, and source data, remain absent. We conduct extensive controlled experiments, generating over one trillion tokens, to identify critical factors in rephrasing web text into synthetic pretraining data. Our results reveal that structured output formats, such as tables, math problems, FAQs, and tutorials, consistently outperform both curated web baselines and prior synthetic methods. Notably, increasing the size of the generator model beyond 1B parameters provides no additional benefit. Our analysis also demonstrates that the selection of the original data used for mixing substantially influences performance. By applying our findings, we develop FinePhrase, a 486-billion-token open dataset of rephrased web text. We show that FinePhrase outperforms all existing synthetic data baselines while reducing generation costs by up to 30 times. We provide the dataset, all prompts, and the generation framework to the research community.

Joel Niklaus, Atsuki Yamaguchi, Michal Štefánik +9

Hugging Face · University of Sheffield, United Kingdom · National Institute of Informatics, Japan

Synthetic Data PretrainingSynthetic Data Generation for Language ModelsSynthetic Data SelectionPrompt Engineering
AskPDF
Apr 14, 2026

Epidemiology of Model Collapse: Modeling Synthetic Data Contamination via Bilayer SIR Dynamics

Training on synthetic data causes model collapse, but existing analyses treat this as single-chain degradation. In reality, the AI ecosystem involves cross-contamination: models ingest synthetic data from other models, produce new synthetic text, and contaminate shared corpora. We propose a bilayer coupled SIR/SIRS framework -- a phenomenological mean-field model treating data corpora and AI models as two interacting populations, each with susceptible, infected, and recovered compartments linked by cross-layer transmission. The SIRS variant (our primary recommendation) incorporates immunity waning, reflecting that filtered corpora and retrained models remain susceptible to re-contamination. We derive the basic reproduction number R0=βDβM/[(γD+μD)(γM+μM)]R_0 = \sqrt{β_D β_M / [(γ_D+μ_D)(γ_M+μ_M)]} via the Next Generation Matrix and apply standard epidemic threshold results to the bilayer system. Illustrative scenario-based calibration from public AI text prevalence data yields supercritical dynamics (R0>1R_0 > 1) across three scenarios; Sobol sensitivity analysis identifies synthetic-text detection as the highest-leverage parameter. A bipartite-network agent-based model confirms mean-field consistency (R2>0.96R^2 > 0.96) for dense networks but degrades under heterogeneity. GPT-2 contamination chain experiments (192 runs across WikiText and Shakespeare) show dose-response degradation and diversity loss qualitatively consistent with the threshold picture. Matched-budget source-diversity experiments (1,088 runs) provide suggestive evidence that multi-source mixing modestly attenuates collapse, but the effect vanishes at lower contamination fractions. Intervention analysis identifies detection-based filtering and herd immunity as the highest-leverage strategies.

Xiangyu Wang

Model CollapseData Contamination in Language ModelsSynthetic Data Pretraining
AskPDF
Apr 10, 2026

SynFlow: Scaling Up LiDAR Scene Flow Estimation with Synthetic Data

Reliable 3D dynamic perception requires models that can anticipate motion beyond predefined categories, yet progress is hindered by the scarcity of dense, high-quality motion annotations. While self-supervision on unlabeled real data offers a path forward, empirical evidence suggests that scaling unlabeled data fails to close the performance gap due to noisy proxy signals. In this paper, we propose learning robust real-world motion priors entirely from scalable simulation. We introduce SynFlow, a data generation pipeline for large-scale synthetic LiDAR scene flow. Unlike prior works that prioritize sensor-specific realism, SynFlow employs a motion-oriented strategy to synthesize diverse kinematic patterns across 4,000 sequences (∼\sim940k frames), termed SynFlow-4k. This represents a 34×34\times scale-up in annotated volume over existing real-world benchmarks. Our experiments demonstrate that SynFlow-4k provides a highly domain-invariant motion prior. In a zero-shot regime, models trained only on our synthetic data generalize across multiple real-world benchmarks, comparable to in-domain supervised baselines on nuScenes and outperforming state-of-the-art methods on TruckScenes by 31.8%. Furthermore, SynFlow-4k serves as a label-efficient foundation: fine-tuning with only 5% of real-world labels surpasses models trained from scratch on the full available budget. We open-source the pipeline and dataset to facilitate research in generalizable 3D motion estimation. More detail can be found at https://kin-zhang.github.io/SynFlow.

Qingwen Zhang, Xiaomeng Zhu, Chenhan Jiang +1

KTH Royal Institute of Technology, Stockholm, Sweden · Hong Kong University of Science and Technology, Hong Kong, China

Synthetic Data PretrainingSynthetic Data GenerationSynthetic-to-Real Domain Adaptation
AskPDF
Apr 8, 2026

WRAP++: Web discoveRy Amplified Pretraining

Synthetic data rephrasing has emerged as a powerful technique for enhancing knowledge acquisition during large language model (LLM) pretraining. However, existing approaches operate at the single-document level, rewriting individual web pages in isolation. This confines synthesized examples to intra-document knowledge, missing cross-document relationships and leaving facts with limited associative context. We propose WRAP++ (Web discoveRy Amplified Pretraining), which amplifies the associative context of factual knowledge by discovering cross-document relationships from web hyperlinks and synthesizing joint QA over each discovered document pair. Concretely, WRAP++ discovers high-confidence relational motifs including dual-links and co-mentions, and synthesizes QA that requires reasoning across both documents. This produces relational knowledge absent from either source document alone, creating diverse entry points to the same facts. Because the number of valid entity pairs grows combinatorially, this discovery-driven synthesis also amplifies data scale far beyond single-document rewriting. Instantiating WRAP++ on Wikipedia, we amplify ~8.4B tokens of raw text into 80B tokens of cross-document QA data. On SimpleQA, OLMo-based models at both 7B and 32B scales trained with WRAP++ substantially outperform single-document approaches and exhibit sustained scaling gains, underscoring the advantage of cross-document knowledge discovery and amplification.

Jiang Zhou, Yunhao Wang, Xing Wu +2

Knowledge Augmentation for Language ModelsSynthetic Data PretrainingSynthetic Data Generation for Language ModelsMulti-Hop QA
AskPDF
Apr 6, 2026

Dynamic Linear Coregionalization for Realistic Synthetic Multivariate Time Series

Synthetic data is essential for training foundation models for time series (FMTS), but most generators assume static correlations, and are typically missing realistic inter-channel dependencies. We introduce DynLMC, a Dynamic Linear Model of Coregionalization, that incorporates time-varying, regime-switching correlations and cross-channel lag structures. Our approach produces synthetic multivariate time series with correlation dynamics that closely resemble real data. Fine-tuning three foundational models on DynLMC-generated data yields consistent zero-shot forecasting improvements across nine benchmarks. Our results demonstrate that modeling dynamic inter-channel correlations enhances FMTS transferability, highlighting the importance of data-centric pretraining.

Annita Vapsi, Penghang Liu, Saheed Obitayo +8

JPMorganChase

Synthetic Data GenerationTime Series GenerationSynthetic Data PretrainingStochastic Linear Dynamical SystemsZero-Shot Time Series Forecasting
AskPDF
Mar 31, 2026

Survival In-Context: Amortized Bayesian Survival Analysis via Prior-Fitted Networks

Survival analysis is crucial for many medical applications, but remains challenging for modern machine learning due to limited data, censoring, and the heterogeneity of tabular covariates. While the prior-fitted paradigm, which relies on pretraining models on large collections of synthetic datasets, has recently facilitated tabular foundation models for classification and regression, its suitability for time-to-event modeling remains unclear. We propose a flexible survival data generation framework that defines a rich survival prior with explicit control over covariates and time-event distributions. Building on this prior, we introduce Survival In-Context (SIC), a prior-fitted in-context learning model for survival analysis that is pretrained exclusively on synthetic data. SIC is trained to approximate Bayesian posterior predictive inference under the synthetic survival prior, enabling individualized survival prediction in a single forward pass, requiring no task-specific training or hyperparameter tuning. Across a broad evaluation on real-world survival datasets, SIC achieves competitive or superior performance compared to classical and deep survival models, particularly in small and medium-sized data regimes, highlighting the promise of a prior-fitted paradigm for survival analysis. The code and pretrained models will be made available upon publication.

Dmitrii Seletkov, Paul Hager, Georgios Kaissis +3

Institute of Diagnostic and Interventional Radiology, Technical University of Munich, Germany · Chair for AI in Healthcare and Medicine, Technical University of Munich, Germany · Hasso Plattner Institute for Digital Engineering, University of Potsdam, Germany +3

Survival AnalysisPrior-Data Fitted NetworksIn-Context LearningAmortized InferenceSynthetic Data Pretraining
AskPDF
Mar 30, 2026

The Surprising Effectiveness of Noise Pretraining for Implicit Neural Representations

The approximation and convergence properties of implicit neural representations (INRs) are known to be highly sensitive to parameter initialization strategies. While several data-driven initialization methods demonstrate significant improvements over standard random sampling, the reasons for their success -- specifically, whether they encode classical statistical signal priors or more complex features -- remain poorly understood. In this study, we explore this phenomenon through a series of experimental analyses leveraging noise pretraining. We pretrain INRs on diverse noise classes (e.g., Gaussian, Dead Leaves, Spectral) and measure their ability to both fit unseen signals and encode priors for an inverse imaging task (denoising). Our analyses on image and video data reveal a surprising finding: simply pretraining on unstructured noise (Uniform, Gaussian) dramatically improves signal fitting capacity compared to all other baselines. However, unstructured noise also yields poor deep image priors for denoising. In contrast, we also find that noise with the classic 1/∣fα∣1/|f^α| spectral structure of natural images achieves an excellent balance of signal fitting and inverse imaging capabilities, performing on par with the best data-driven initialization methods. This finding enables more efficient INR training in applications lacking sufficient prior domain-specific data. For more details, visit project page at https://kushalvyas.github.io/noisepretraining.html

Kushal Vyas, Alper Kayabasi, Daniel Kim +3

Rice University · University of California, Riverside

Implicit Neural RepresentationsNeural Network InitializationSynthetic Data PretrainingSelf-Supervised Pre-Training
AskPDF
Mar 26, 2026

Comparing Natural and Synthetic Structured Data: A Study of the Passive Verb Alternation in French and Italian

This study compares the impact of natural and synthetic data on training and evaluating large language models (LLMs), using the case of passive verb alternation in French and Italian. We use Blackbird Language Matrices (BLMs), structured datasets designed to probe linguistic knowledge of underlying patterns across sentence sets. We compare structured templates instantiated with natural sentences extracted from Universal Dependencies to structured templates of synthetic sentences. Experiments show that while models achieve ceiling performance when trained and tested on synthetic datasets, they do not reliably generalize to natural sentences. In contrast, models trained on natural data exhibit robust performance across both natural and synthetic test suites, demonstrating their superior ability to capture abstract linguistic patterns. These results corroborate the value of natural data and of structured set ups in linguistic evaluation for probing LLMs' syntactic and semantic knowledge.

Giuseppe Samo, Paola Merlo

IDIAP Research Institute · University of Geneva

Synthetic Data EvaluationSynthetic Data PretrainingMultilingual Language Model Evaluation
AskPDF
Mar 25, 2026

CORA: Generalizable coronary artery disease assessment and risk stratification from coronary CT angiography using pathology-centric representation learning

Coronary artery disease, a leading cause of cardiovascular mortality worldwide, can be assessed non-invasively by coronary computed tomography angiography (CCTA). Although deep learning has advanced automated CCTA analysis, clinical translation remains constrained by the scarcity of expert-annotated data and by the spatial sparsity of coronary pathology, which occupies only a small fraction of each scan. Widely used label-free pretraining strategies, such as masked image modeling and contrastive learning, optimize for global anatomical reconstruction and tend to under-represent these tiny localized pathological features. Here we present CORA, an annotation-efficient model for comprehensive coronary artery disease assessment. Rather than reconstructing background anatomy, CORA learns from volumetric CCTA through a synthesis-driven self-supervised strategy: an anatomy-guided engine inserts diverse synthetic calcified and non-calcified lesions into unlabeled scans, reframing pretraining as an abnormality-detection task that biases representation learning toward clinically relevant disease features. We pretrained CORA on 10,138 unlabeled CCTA volumes and evaluated it across datasets from nine independent hospitals. Across plaque characterization, stenosis detection, and coronary artery segmentation, CORA consistently outperformed strong self-supervised pretraining baselines, with the largest gains on external multi-center data, indicating robust generalization under distributional shift. Coupling the imaging encoder with structured clinical variables further enabled near-term major adverse cardiac event (MACE) risk stratification. Our results show that pathology-centric, synthesis-driven pretraining is an effective and scalable strategy for annotation-efficient coronary artery disease assessment from CCTA.

Jinkui Hao, Gorkem Durak, Halil Ertugrul Aktas +4

Department of Radiology, Northwestern University, Chicago, IL, USA · Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA · Department of Cardiology, Northwestern University, Chicago, IL, USA +1

Synthetic Data PretrainingCoronary AngiographyMedical DiagnosisMedical Image Anomaly DetectionMedical Image SegmentationCoronary Artery Segmentation+3
AskPDF
Mar 19, 2026

Data-efficient pre-training by scaling synthetic megadocs

Synthetic data augmentation has emerged as a promising solution when pre-training is constrained by data rather than compute. We study how to design synthetic data algorithms that achieve better loss scaling: not only lowering loss at finite compute but especially as compute approaches infinity. We first show that pre-training on web data mixed with synthetically generated rephrases improves i.i.d. validation loss on the web data, despite the synthetic data coming from an entirely different distribution. With optimal mixing and epoching, loss and benchmark accuracy improve without overfitting as the number of synthetic generations grows, plateauing near 1.48×1.48\times data efficiency at 32 rephrases per document. We find even better loss scaling under a new perspective: synthetic generations from the same document can form a single substantially longer megadocument instead of many short documents. We show two ways to construct megadocs: stitching synthetic rephrases from the same web document or stretching a document by inserting rationales. Both methods improve i.i.d. loss, downstream benchmarks, and especially long-context loss relative to simple rephrasing, increasing data efficiency from 1.48×1.48\times to 1.80×1.80\times at 3232 generations per document. Importantly, the improvement of megadocs over simple rephrasing widens as more synthetic data is generated. Our results show how to design synthetic data algorithms that benefit more from increasing compute when data-constrained.

Konwoo Kim, Suhas Kotha, Yejin Choi +3

Stanford University

Language Model PretrainingSynthetic Data PretrainingSynthetic Data AugmentationSynthetic Data Generation for Language ModelsLanguage Model Scaling Laws
AskPDF
Mar 18, 2026

Continually self-improving AI

Modern language model-based AI systems are remarkably powerful, yet their capabilities remain fundamentally capped by their human creators in three key ways. First, although a model's weights can be updated via fine-tuning, acquiring new knowledge from small, specialized corpora after pretraining remains highly data-inefficient. Second, the training of these systems relies heavily on finite, human-generated data from across history. Third, the pipelines used to train AI models are confined by the algorithms that human researchers can discover and explore. This thesis takes a small step toward overcoming these inherent limitations, presenting three chapters aimed at breaking these dependencies to create continually self-improving AI. First, to overcome this data-efficiency barrier in knowledge acquisition, we propose a synthetic data approach that diversifies and amplifies small corpora into rich knowledge representations, enabling a model to effectively update its parameters from limited source material. Second, to reduce reliance on human data, we show that given a fixed amount of such data, the model can self-generate synthetic data to bootstrap its fundamental pretraining capabilities without distillation from any off-the-shelf, instruction-tuned LM. Finally, to transcend human-engineered training paradigms, we demonstrate that by scaling search during test time over the space of algorithms, AI can search over a larger space of learning algorithm configurations than human researchers can explore manually.

Zitong Yang

Department of Statistics, Stanford University

Language Model Self-ImprovementSelf-Training for Language ModelsKnowledge Augmentation for Language ModelsSynthetic Data Generation for Language ModelsSynthetic Data Pretraining
AskPDF
Mar 9, 2026

Training Language Models via Neural Cellular Automata

Pre-training is crucial for large language models (LLMs), as it is when most representations and capabilities are acquired. However, natural language pre-training has problems: high-quality text is finite, it contains human biases, and it entangles knowledge with reasoning. This raises a fundamental question: is natural language the only path to intelligence? We propose using neural cellular automata (NCA) to generate synthetic, non-linguistic data for pre-pre-training LLMs--training on synthetic-then-natural language. NCA data exhibits rich spatiotemporal structure and statistics resembling natural language while being controllable and cheap to generate at scale. We find that pre-pre-training on only 164M NCA tokens improves downstream language modeling by up to 6% and accelerates convergence by up to 1.6x. Surprisingly, this even outperforms pre-pre-training on 1.6B tokens of natural language from Common Crawl with more compute. These gains also transfer to reasoning benchmarks, including GSM8K, HumanEval, and BigBench-Lite. Investigating what drives transfer, we find that attention layers are the most transferable, and that optimal NCA complexity varies by domain: code benefits from simpler dynamics, while math and web text favor more complex ones. These results enable systematic tuning of the synthetic distribution to target domains. More broadly, our work opens a path toward more efficient models with fully synthetic pre-training.

Dan Lee, Seungwook Han, Akarsh Kumar +1

Independent Contributor · MIT · Improbable AI Lab

Language Model PretrainingSynthetic Data PretrainingSynthetic Data Generation for Language ModelsNeural Cellular Automata
AskPDF
Mar 5, 2026

Engineering Regression Without Real-Data Training: Domain Adaptation for Tabular Foundation Models Using Multi-Dataset Embeddings

Predictive modeling in engineering applications has long been dominated by bespoke models and small, siloed tabular datasets, limiting the applicability of large-scale learning approaches. Despite recent progress in tabular foundation models, the resulting synthetic training distributions used for pre-training may not reflect the statistical structure of engineering data, limiting transfer to engineering regression. We introduce TREDBench, a curated collection of 83 real-world tabular regression datasets with expert engineering/non-engineering labels, and use TabPFN 2.5's dataset-level embedding to study domain structure in a common representation space. We find that engineering datasets are partially distinguishable from non-engineering datasets, while standard procedurally generated datasets are highly distinguishable from engineering datasets, revealing a substantial synthetic-real domain gap. To bridge this gap without training on real engineering samples, we propose an embedding-guided synthetic data curation method: we generate and identify "engineering-like" synthetic datasets, and perform continued pre-training of TabPFN 2.5 using only the selected synthetic tasks. Across 35 engineering regression datasets, this synthetic-only adaptation improves predictive accuracy and data efficiency, outperforming TabPFN 2.5 on 29/35 datasets and AutoGluon on 27/35, with mean multiplicative data-efficiency gains of 1.75x and 4.44x, respectively. More broadly, our results indicate that principled synthetic data curation can convert procedural generators into domain-relevant "data engines," enabling foundation models to improve in data-sparse scientific and industrial domains where real data collection is the primary bottleneck.

Lyle Regenwetter, Rosen Yu, Cyril Picard +1

Massachusetts Institute of Technology

Synthetic Data PretrainingSynthetic-to-Real Domain AdaptationTabular Foundation ModelsSynthetic Data Selection
AskPDF
Mar 4, 2026

Relational In-Context Learning via Synthetic Pre-training with Structural Prior

Relational Databases (RDBs) are the backbone of modern business, yet they lack foundation models comparable to those in text or vision. A key obstacle is that high-quality RDBs are private, scarce, and structurally heterogeneous, making internet-scale pre-training infeasible. To overcome this data scarcity, we introduce RDB-PFN, the first relational foundation model trained purely via synthetic data. Inspired by Prior-Data Fitted Networks (PFNs), where synthetic data generated from Structural Causal Models (SCMs) enables reasoning on single tables, we design a Relational Prior Generator to create an infinite stream of diverse RDBs from scratch. Pre-training on over 2 million synthetic single-table and relational tasks, RDB-PFN learns to adapt to any new database instantly via genuine in-context learning. Experiments show that RDB-PFN achieves strong few-shot performance on 19 real-world relational prediction tasks, outperforming state-of-the-art tabular foundation models evaluated on the same DFS-linearized inputs, while using a lightweight architecture and fast inference. The code is available at https://github.com/MuLabPKU/RDBPFN.

Yanbo Wang, Jiaxuan You, Chuan Shi +1

Institute for Artificial Intelligence, Peking University · University of Illinois at Urbana-Champaign · Institute of Computing Technology, Beijing University of Post and Telecommunication +1

Relational ICLRelational Foundation ModelsRelational DatabasesIn-Context LearningSynthetic Data PretrainingSynthetic Data Generation
AskPDF
Mar 2, 2026

Reasoning Core: A Scalable Procedural Data Generation Suite for Symbolic Pre-training and Post-Training

Training on verifiable symbolic data is a promising way to expand the reasoning frontier of language models beyond what standard pre-training corpora provide. Yet existing procedural generators often rely on fixed puzzles or templates and do not deliver the distributional breadth needed at scale. We introduce Reasoning Core, a scalable suite that procedurally generates verifiable symbolic reasoning data across core formal domains: PDDL planning over randomized domains, first-order logic with equality, context-free grammar parsing and generation, causal reasoning over random Bayesian networks, and systems of equations. Each task is paired with an external solver for rigorous verification and admits continuous difficulty control for curriculum design. Examples can optionally include solver-derived reasoning traces, enabling supervised training from the earliest pre-training stages, and the same interface provides verifiable reward functions for reinforcement learning. Our experiments show that mixing Reasoning Core data into pre-training improves downstream reasoning while preserving, or slightly improving, language modeling quality. Zero-shot evaluations confirm these tasks challenge frontier models such as GPT-5. The code and data are publicly available under the MIT license.

Valentin Lacombe, Valentin Quesnel, Damien Sileo

Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 - CRIStAL, F-59000 Lille, France

Synthetic Data Generation for Language ModelsSynthetic Data PretrainingSymbolic Reasoning
AskPDF
Mar 1, 2026

Fake It Right: Injecting Anatomical Logic into Synthetic Supervised Pre-training for Medical Segmentation

Vision Transformers (ViTs) excel in 3D medical segmentation but require massive annotated datasets. While Self-Supervised Learning (SSL) mitigates this using unlabeled data, it still faces strict privacy and logistical barriers. Formula-Driven Supervised Learning (FDSL) offers a privacy-preserving alternative by pre-training on synthetic mathematical primitives. However, a critical semantic gap limits its efficacy: generic shapes lack the morphological fidelity, fixed spatial layouts, and inter-organ relationships of real anatomy, preventing models from learning essential global structural priors. To bridge this gap, we propose an Anatomy-Informed Synthetic Supervised Pre-training framework unifying FDSL's infinite scalability with anatomical realism. We replace basic primitives with a lightweight shape bank with de-identified, label-only segmentation masks from 5 subjects. Furthermore, we introduce a structure-aware sequential placement strategy to govern the patch synthesis process. Instead of random placement, we enforce physiological plausibility using spatial anchors for correct localization and a topological graph to manage inter-organ interactions (e.g., preventing impossible overlaps). Extensive experiments on BTCV and MSD datasets demonstrate that our method significantly outperforms state-of-the-art FDSL baselines and SSL methods by 1.74% and up to 1.66%, while exhibiting a robust scaling effect where performance improves with increased synthetic data volume. This provides a data-efficient, privacy-compliant solution for medical segmentation. The code will be made publicly available upon acceptance.

Jiaqi Tang, Mengyan Zheng, Shu Zhang +2

Peking University · Beihang University · Deepwise Ltd.

Medical Image SegmentationSynthetic Data Pretraining3D Medical Image Segmentation3D ViTs
AskPDF
Feb 23, 2026

GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training

Neural simulators promise efficient surrogates for physics simulation, but scaling them is bottlenecked by the prohibitive cost of generating high-fidelity training data. Pre-training on abundant off-the-shelf geometries offers a natural alternative, yet faces a fundamental gap: supervision on static geometry alone ignores dynamics and can lead to negative transfer on physics tasks. We present GeoPT, a unified pre-trained model for general physics simulation based on lifted geometric pre-training. The core idea is to augment geometry with synthetic dynamics, enabling dynamics-aware self-supervision without physics labels. Pre-trained on over one million samples, GeoPT consistently improves industrial-fidelity benchmarks spanning fluid mechanics for cars, aircraft, and ships, and solid mechanics in crash simulation, reducing labeled data requirements by 20-60% and accelerating convergence by 2×\times. These results show that lifting with synthetic dynamics bridges the geometry-physics gap, unlocking a scalable path for neural simulation and potentially beyond. Code is available at https://github.com/Physics-Scaling/GeoPT.

Haixu Wu, Minghao Guo, Zongyi Li +4

MIT CSAIL · Tsinghua University

Physics-Based SimulationNeural Surrogate ModelingSynthetic Data PretrainingSelf-Supervised Pre-Training
AskPDF
Feb 19, 2026

MantisV2: Closing the Zero-Shot Gap in Time Series Classification with Synthetic Data and Test-Time Strategies

Developing foundation models for time series classification is of high practical relevance, as such models can serve as universal feature extractors for diverse downstream tasks. Although early models such as Mantis have shown the promise of this approach, a substantial performance gap remained between frozen and fine-tuned encoders. In this work, we introduce methods that significantly strengthen zero-shot feature extraction for time series. First, we introduce Mantis+, a variant of Mantis pre-trained entirely on synthetic time series. Second, through controlled ablation studies, we refine the architecture and obtain MantisV2, an improved and more lightweight encoder. Third, we propose an enhanced test-time methodology that leverages intermediate-layer representations and refines output-token aggregation. In addition, we show that performance can be further improved via self-ensembling and cross-model embedding fusion. Extensive experiments on UCR, UEA, Human Activity Recognition (HAR) benchmarks, and EEG datasets show that MantisV2 and Mantis+ consistently outperform prior time series foundation models, achieving state-of-the-art zero-shot performance.

Vasilii Feofanov, Songkang Wen, Jianfeng Zhang +2

Huawei Noah’s Ark Lab

Time Series ClassificationTime-Series Foundation ModelsSynthetic Data Pretraining
AskPDF
Feb 16, 2026

Universal priors: solving empirical Bayes via Bayesian inference and pretraining

We theoretically justify the recent empirical finding of [Teh et al., 2025] that a transformer pretrained on synthetically generated data achieves strong performance on empirical Bayes (EB) problems. We take an indirect approach to this question: rather than analyzing the model architecture or training dynamics, we ask why a pretrained Bayes estimator, trained under a prespecified training distribution, can adapt to arbitrary test distributions. Focusing on Poisson EB problems, we identify the existence of universal priors such that training under these priors yields a near-optimal regret bound of O~(1n)\widetilde{O}(\frac{1}{n}) uniformly over all test distributions. Our analysis leverages the classical phenomenon of posterior contraction in Bayesian statistics, showing that the pretrained transformer adapts to unknown test distributions precisely through posterior contraction. This perspective also explains the phenomenon of length generalization, in which the test sequence length exceeds the training length, as the model performs Bayesian inference using a generalized posterior.

Nick Cannella, Anzo Teh, Yanjun Han +1

Courant Institute of Mathematical Sciences, NYU, New York City, NY · Department of EECS, MIT, Cambridge, MA · Courant Institute of Mathematical Sciences and Center for Data Science, NYU, New York City, NY

Synthetic Data PretrainingEmpirical BayesBayesian InferencePosterior Contraction
AskPDF
Feb 11, 2026

TabICLv2: A better, faster, scalable, and open tabular foundation model

Tabular foundation models, such as TabPFNv2 and TabICL, have recently dethroned gradient-boosted trees at the top of predictive benchmarks, demonstrating the value of in-context learning for tabular data. We introduce TabICLv2, a new state-of-the-art foundation model for regression and classification built on three pillars: (1) a novel synthetic data generation engine designed for high pretraining diversity; (2) various architectural innovations, including a new scalable softmax in attention improving generalization to larger datasets without prohibitive long-sequence pretraining; and (3) optimized pretraining protocols, notably replacing AdamW with the Muon optimizer. On the TabArena and TALENT benchmarks, TabICLv2 without any tuning surpasses the performance of the current state of the art, RealTabPFN-2.5 (hyperparameter-tuned, ensembled, and fine-tuned on real data). With only moderate pretraining compute, TabICLv2 generalizes effectively to million-scale datasets under 50 GB GPU memory while being markedly faster than RealTabPFN-2.5. We provide extensive ablation studies to quantify these contributions and foster open research by releasing code for inference, pretraining, and synthetic data generation at https://github.com/soda-inria/tabicl.

Jingang Qu, David Holzmüller, Gaël Varoquaux +1

SODA Team, INRIA Saclay, Palaiseau, France · Probabl, France

Synthetic Data PretrainingTabular Foundation ModelsSynthetic Tabular Data GenerationTabular ICLTabular Prediction
AskPDF
Feb 3, 2026

PluRel: Synthetic Data unlocks Scaling Laws for Relational Foundation Models

Relational Foundation Models (RFMs) facilitate data-driven decision-making by learning from complex multi-table databases. However, the diverse relational databases needed to train such models are rarely public due to privacy constraints. While there are methods to generate synthetic tabular data of arbitrary size, incorporating schema structure and primary-foreign key connectivity for multi-table generation remains challenging. Here we introduce PLUREL, a framework to synthesize multi-tabular relational databases from scratch. In a step-by-step fashion, PLUREL models (1) schemas with directed graphs, (2) inter-table primary-foreign key connectivity with bipartite graphs, and, (3) feature distributions in tables via conditional causal mechanisms. The design space across these stages supports the synthesis of a wide range of diverse databases, while being computationally lightweight. Using PLUREL, we observe for the first time that (1) RFM pretraining loss exhibits power-law scaling with the number of synthetic databases and total pretraining tokens, (2) scaling the number of synthetic databases improves generalization to real databases, and (3) synthetic pretraining yields strong base models for continued pretraining on real databases. Overall, our framework and results position synthetic data scaling as a promising paradigm for RFMs.

Vignesh Kothapalli, Rishabh Ranjan, Valter Hudovernik +4

1Stanford University · 2Kumo AI · 3SAP Labs LLC

Relational Foundation ModelsSynthetic Data PretrainingSynthetic Data GenerationSynthetic Tabular Data GenerationRelational Deep LearningRelational Databases
AskPDF
Jan 29, 2026

FineInstructions: Scaling Synthetic Instructions to Pre-Training Scale

Due to limited supervised training data, large language models (LLMs) are typically pre-trained via a self-supervised "predict the next word" objective on a vast amount of unstructured text data. To make the resulting model useful to users, it is further trained on a far smaller amount of "instruction-tuning" data comprised of supervised training examples of instructions and responses. To overcome the limited amount of supervised data, we propose a procedure that can transform the knowledge in internet-scale pre-training documents into billions of synthetic instruction and answer training pairs. The resulting dataset, called FineInstructions, uses ~18M instruction templates created from real user-written queries and prompts. These instruction templates are matched to and instantiated with human-written source documents from unstructured pre-training corpora. With "supervised" synthetic training data generated at this scale, an LLM can be pre-trained from scratch solely with the instruction-tuning objective, which is far more in-distribution with the expected downstream usage of LLMs (responding to user prompts). We conduct controlled token-for-token training experiments and find pre-training on FineInstructions outperforms standard pre-training and other proposed synthetic pre-training techniques on standard benchmarks measuring free-form response quality. Our resources can be found at https://huggingface.co/fineinstructions .

Ajay Patel, Colin Raffel, Chris Callison-Burch

University of Pennsylvania · University of Toronto · Vector Institute +1

Language Model PretrainingSynthetic Data PretrainingSynthetic Data GenerationSynthetic Instruction GenerationInstruction TuningSynthetic Data Generation for Language Models
AskPDF
Jan 28, 2026

ECGFlowCMR: Pretraining with ECG-Generated Cine CMR Helps Cardiac Disease Classification and Phenotype Prediction

Cardiac Magnetic Resonance (CMR) imaging provides a comprehensive assessment of cardiac structure and function but remains constrained by high acquisition costs and reliance on expert annotations, limiting the availability of large-scale labeled datasets. In contrast, electrocardiograms (ECGs) are inexpensive, widely accessible, and offer a promising modality for conditioning the generative synthesis of cine CMR. To this end, we propose ECGFlowCMR, a novel ECG-to-CMR generative framework that integrates a Phase-Aware Masked Autoencoder (PA-MAE) and an Anatomy-Motion Disentangled Flow (AMDF) to address two fundamental challenges: (1) the cross-modal temporal mismatch between multi-beat ECG recordings and single-cycle CMR sequences, and (2) the anatomical observability gap due to the limited structural information inherent in ECGs. Extensive experiments on the UK Biobank and a proprietary clinical dataset demonstrate that ECGFlowCMR can generate realistic cine CMR sequences from ECG inputs, enabling scalable pretraining and improving performance on downstream cardiac disease classification and phenotype prediction tasks.

Xiaocheng Fang, Zhengyao Ding, Guangkun Nie +9

State Key Laboratory of General Artificial Intelligence, Peking University · School of Intelligence Science and Technology, Peking University, Beijing, China · Polytechnic Institute of Zhejiang University, Zhejiang University, Hangzhou, China +4

Synthetic Data PretrainingMedical DiagnosisClinical PredictionMedical Image SynthesisConditional Generative ModelingCardiac MRI
AskPDF
Dec 28, 2025

Long-Range Distillation: Distilling 10,000 Years of Simulated Climate into Long Timestep AI Weather Models

Accurate long-range weather forecasting remains a major challenge for AI models, both because errors accumulate over autoregressive rollouts and because reanalysis training datasets provide limited sampling of the slow modes of variability underpinning predictability. Most AI and all numerical weather prediction models are autoregressive, repeatedly applying short-lead forecasts to reach subseasonal-to-seasonal (S2S) or seasonal lead times, often causing instability and miscalibration. Long-timestep probabilistic models that generate long-range forecasts in a single step offer an alternative, but training on the 40-year reanalysis record leads to overfitting, suggesting orders of magnitude more data are required. We introduce long-range distillation, whereby a long-timestep probabilistic student'' model is trained to forecast directly at long-range using a huge synthetic dataset generated by an autoregressive teacher'' model. Using the Deep Learning Earth System Model (DLESyM) as teacher, we generate over 10,000 years of simulated climate to train student models across a range of timescales. In perfect-model experiments, the distilled models outperform climatology and approach their teacher's skill while replacing hundreds of autoregressive steps with a single timestep. In the real world, both a diffusion model student fine-tuned on ERA5 and a linear inverse model student with mixed DLESyM-ERA5 training achieve S2S forecast skill competitive with the ECMWF IFS ensemble, with synthetic pre-training providing improvements beyond ERA5-only training. This demonstrates the benefit of synthetic pre-training for both deep learning and classical data-driven methods, and represents the first demonstration that real-world skill can be improved by training on AI-generated synthetic data in volumes far exceeding the observational record.

Scott A. Martin, Noah Brenowitz, Dale Durran +1

NVIDIA Research, Santa Clara, CA, USA · School of Oceanography, University of Washington, Seattle, WA, USA · Department of Atmospheric & Climate Science, University of Washington, Seattle, WA, USA

Weather ForecastingSpatiotemporal ForecastingKnowledge DistillationSynthetic Data PretrainingProbabilistic Forecasting
AskPDF
Nov 6, 2025

Towards Scalable Meta-Learning of near-optimal Interpretable Models via Synthetic Model Generations

Decision trees are widely used in high-stakes fields like finance and healthcare due to their interpretability. This work introduces an efficient, scalable method for generating synthetic pre-training data to enable meta-learning of decision trees. Our approach samples near-optimal decision trees synthetically, creating large-scale, realistic datasets. Using the MetaTree transformer architecture, we demonstrate that this method achieves performance comparable to pre-training on real-world data or with computationally expensive optimal decision trees. This strategy significantly reduces computational costs, enhances data generation flexibility, and paves the way for scalable and efficient meta-learning of interpretable decision tree models.

Kyaw Hpone Myint, Zhe Wu, Alexandre G. R. Day +1

Capital One

Decision Tree LearningMeta-LearningInterpretable MLSynthetic Data PretrainingSynthetic Data Generation
AskPDF
Oct 18, 2025

Escaping Model Collapse via Synthetic Data Verification: Near-term Improvements and Long-term Convergence

Synthetic data has been increasingly used to train frontier generative models. However, recent studies raise key concerns that iteratively retraining a generative model on its self-generated synthetic data may keep deteriorating model performance, a phenomenon often coined model collapse. In this paper, we investigate ways to modify the synthetic retraining process to avoid model collapse, and even possibly help reverse the trend from collapse to improvement. Our key finding is that by injecting information through an external synthetic data verifier, whether a human or a better model, synthetic retraining will not cause model collapse. Specifically, we situate our theoretical analysis in the fundamental linear regression setting, showing that verifier-guided retraining can yield near-term improvements, but ultimately drives the parameter estimate to the verifier's "knowledge center" in the long run. Our theory further predicts that, unless the verifier is perfectly reliable, these early gains will plateau and may even reverse. Indeed, our experiments across linear regression, Variational Autoencoders (VAEs) trained on MNIST, and fining-tuning SmolLM2-135M on the XSUM task confirm these theoretical insights.

Bingji Yi, Qiyuan Liu, Yuwei Cheng +1

Independent Researcher. Work done while visiting UChicago CS. · Work done while visiting UChicago CS. · Department of Statistics, University of Chicago. +1

Synthetic Data PretrainingModel CollapseSynthetic Data Generation for Language ModelsSynthetic Data Evaluation
AskPDF
Sep 25, 2025

GraphPFN: A Prior-Data Fitted Graph Foundation Model

Graph foundation models face several fundamental challenges including transferability across diverse domains and data scarcity, which calls into question the very feasibility of creating such models. However, despite similar challenges, the tabular domain has recently witnessed the emergence of the first successful foundation models such as TabPFN. These models are based on the prior-data fitted networks (PFN) framework, in which models are pretrained on carefully designed synthetic datasets to make predictions in an in-context learning setting. Recently, G2T-FM, a framework that converts graph node-level tasks into tabular tasks, has made the first step towards adopting PFNs for graphs, yet it is limited to hand-crafted features and was never pretrained on graph data. In this work, we make the next step by proposing GraphPFN, a PFN-based model designed and pretrained specifically for graph node-level tasks. Following the PFN framework, we first design a prior distribution of synthetic attributed graphs by using a novel combination of multi-level stochastic block models and a preferential attachment process for structure generation and graph-aware structured causal models for attribute generation. Then, we augment the tabular foundation model LimiX with attention-based graph neighborhood aggregation layers and train it on millions of synthetic graphs sampled from our prior. On diverse real-world graph datasets with node-level tasks, GraphPFN achieves state-of-the-art results in both in-context learning and finetuning regimes, outperforming G2T-FM, prior GFMs, and task-specific GNNs trained from scratch. More broadly, GraphPFN shows the potential of PFN-based models for building graph foundation models.

Dmitry Eremeev, Oleg Platonov, Gleb Bazhenov +2

1HSE University · 2Yandex Research

Synthetic Data PretrainingIn-Context LearningGraph Foundation ModelsPrior-Data Fitted NetworksNode Classification
AskPDF