# Self-Correcting Self-Consuming Loops for Generative Model Training

arXiv: https://arxiv.org/abs/2402.07087 · HTML: https://mlsanity.com/paper/2402.07087/

## Key facts

- Cited by 37 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 34 references that are in the map.
- First posted on arXiv 11 Feb 2024.

## Builds on (most cited first, 24 of 24)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding](https://mlsanity.com/paper/2205.11487/) (2022)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [Expressive Body Capture: 3D Hands, Face, and Body from a Single Image](https://mlsanity.com/paper/1904.05866/) (2019)
- [Human Motion Diffusion Model](https://mlsanity.com/paper/2209.14916/) (2022)
- [Self-critiquing models for assisting human evaluators](https://mlsanity.com/paper/2206.05802/) (2022)
- [PhysDiff: Physics-Guided Human Motion Diffusion Model](https://mlsanity.com/paper/2212.02500/) (2022)
- [The Curse of Recursion: Training on Generated Data Makes Models Forget](https://mlsanity.com/paper/2305.17493/) (2023)
- [Performative Prediction](https://mlsanity.com/paper/2002.06673/) (2020)
- [Synthetic Data from Diffusion Models Improves ImageNet Classification](https://mlsanity.com/paper/2304.08466/) (2023)
- [Is synthetic data from generative models ready for image recognition?](https://mlsanity.com/paper/2210.07574/) (2022)
- [Generating Sequences by Learning to Self-Correct](https://mlsanity.com/paper/2211.00053/) (2022)
- [Self-Consuming Generative Models Go MAD](https://mlsanity.com/paper/2307.01850/) (2023)
- [InterDiff: Generating 3D Human-Object Interactions with Physics-Informed Diffusion](https://mlsanity.com/paper/2308.16905/) (2023)
- [Self-correcting LLM-controlled Diffusion Models](https://mlsanity.com/paper/2311.16090/) (2023)
- [Generative Models as a Data Source for Multiview Representation Learning](https://mlsanity.com/paper/2106.05258/) (2021)
- [Dynamics-Regulated Kinematic Policy for Egocentric Pose Estimation](https://mlsanity.com/paper/2106.05969/) (2021)
- [On the Stability of Iterative Retraining of Generative Models on their own Data](https://mlsanity.com/paper/2310.00429/) (2023)
- [Learning Vision from Models Rivals Learning Vision from Data](https://mlsanity.com/paper/2312.17742/) (2023)
- [MoVi: A Large Multipurpose Motion and Video Dataset](https://mlsanity.com/paper/2003.01888/) (2020)
- [Towards Understanding the Interplay of Generative Artificial Intelligence and the Internet](https://mlsanity.com/paper/2306.06130/) (2023)
- [Generalization Error in Deep Learning](https://mlsanity.com/paper/1808.01174/) (2018)

## Built on by (most cited first, 30 of 37)

- [Model Collapse Demystified: The Case of Regression](https://mlsanity.com/paper/2402.07712/) (2024)
- [Can Large Reasoning Models Self-Train?](https://mlsanity.com/paper/2505.21444/) (2025)
- [Strong Model Collapse](https://mlsanity.com/paper/2410.04840/) (2024)
- [Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World](https://mlsanity.com/paper/2410.16713/) (2024)
- [Beyond Model Collapse: Scaling Up with Synthesized Data Requires Verification](https://mlsanity.com/paper/2406.07515/) (2024)
- [Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferences](https://mlsanity.com/paper/2407.09499/) (2024)
- [Self-Improving Diffusion Models with Synthetic Data](https://mlsanity.com/paper/2408.16333/) (2024)
- [Position: Model Collapse Does Not Mean What You Think](https://mlsanity.com/paper/2503.03150/) (2025)
- [Self-Improving Transformers Overcome Easy-to-Hard and Length Generalization Challenges](https://mlsanity.com/paper/2502.01612/) (2025)
- [Mind the Gap: Examining the Self-Improvement Capabilities of Large Language Models](https://mlsanity.com/paper/2412.02674/) (2024)
- [A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops](https://mlsanity.com/paper/2502.18865/) (2025)
- [Model Collapse in the Self-Consuming Chain of Diffusion Finetuning: A Novel Perspective from Quantitative Trait Modeling](https://mlsanity.com/paper/2407.17493/) (2024)
- [Escaping Collapse: The Strength of Weak Data for Large Language Model Training](https://mlsanity.com/paper/2502.08924/) (2025)
- [Self-Consuming Generative Models with Adversarially Curated Data](https://mlsanity.com/paper/2505.09768/) (2025)
- [Why Less is More (Sometimes): A Theory of Data Curation](https://mlsanity.com/paper/2511.03492/) (2025)
- [ModSkill: Physical Character Skill Modularization](https://mlsanity.com/paper/2502.14140/) (2025)
- [LLM Web Dynamics: Tracing Model Collapse in a Network of LLMs](https://mlsanity.com/paper/2506.15690/) (2025)
- [ACDC: Autoregressive Coherent Multimodal Generation using Diffusion Correction](https://mlsanity.com/paper/2410.04721/) (2024)
- [Spend Wisely: Maximizing Post-Training Gains in Iterative Synthetic Data Bootstrapping](https://mlsanity.com/paper/2501.18962/) (2025)
- [Convergence Dynamics and Stabilization Strategies of Co-Evolving Generative Models](https://mlsanity.com/paper/2503.08117/) (2025)
- [SimDiff: Simulator-constrained Diffusion Model for Physically Plausible Motion Generation](https://mlsanity.com/paper/2509.20927/) (2025)
- [Observations and Remedies for Large Language Model Bias in Self-Consuming Performative Loop](https://mlsanity.com/paper/2601.05184/) (2026)
- [Ambient Dataloops: Generative Models for Dataset Refinement](https://mlsanity.com/paper/2601.15417/) (2026)
- [When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs](https://mlsanity.com/paper/2606.28438/) (2026)
- [Large Language Models are Shannon Lossy Compressors Not Solomonoff Induction Estimators: Self-improvement and Singularity Are Not Near Without Symbolic Model Synthesis](https://mlsanity.com/paper/2601.05280/) (2026)
- [Drift and selection in LLM text ecosystems](https://mlsanity.com/paper/2604.08554/) (2026)
- [Beyond MoCap: Scaling Motion Tokenizers with Synthetic Human Motion for Generative Modeling](https://mlsanity.com/paper/2606.27547/) (2026)
- [Skeleton2Stage: Reward-Guided Fine-Tuning for Physically Plausible Dance Generation](https://mlsanity.com/paper/2602.13778/) (2026)
- [A Markovian View of Iterative-Feedback Loops in Image Generative Models: Neural Resonance and Model Collapse](https://mlsanity.com/paper/2602.19033/) (2026)
- [The Oligarch Barely Steers Model Collapse in Multi-Model Ecosystems](https://mlsanity.com/paper/2609.11146/) (2026)

## Abstract (arXiv)

As synthetic data becomes higher quality and proliferates on the internet, machine learning models are increasingly trained on a mix of human- and machine-generated data. Despite the successful stories of using synthetic data for representation learning, using synthetic data for generative model training creates "self-consuming loops" which may lead to training instability or even collapse, unless certain conditions are met. Our paper aims to stabilize self-consuming generative model training. Our theoretical results demonstrate that by introducing an idealized correction function, which maps a data point to be more likely under the true data distribution, self-consuming loops can be made exponentially more stable. We then propose self-correction functions, which rely on expert knowledge (e.g. the laws of physics programmed in a simulator), and aim to approximate the idealized corrector automatically and at scale. We empirically validate the effectiveness of self-correcting self-consuming loops on the challenging human motion synthesis task, and observe that it successfully avoids model collapse, even when the ratio of synthetic data to real data is as high as 100%.