No article text was recorded for this version.
recursive training degradation
Version 2 of 2415 papers736 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Towards Understanding the Interplay of Generative Artificial Intelligence and the Internet
- Blog: Nepotistically Trained Generative Image Models Collapse
- Blog: AI models collapse when trained on recursively generated data
- Blog: Self-Consuming Generative Models go MAD
- Blog: The Curse of Recursion: Training on Generated Data Makes Models Forget
- Blog: Nepotistically Trained Generative-AI Models Collapse
- Blog: Self-Consuming Generative Models Go MAD
- Blog: Self-Consuming Generative Models Go MAD
- Blog: A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective
- Blog: Accumulating Data Avoids Model Collapse
- Blog: How Recursive Training Collapses and What Can Be Done About It
- Blog: Rate of Model Collapse in Recursive Training
- Blog: On the caveats of AI autophagy
- Blog: Combining Generative Artificial Intelligence (AI) and the Internet: Heading towards Evolution or Degradation?
- Blog: Large Language Models as Search Engines: Societal Challenges
- Blog: Investigating Shifts in GAN Output-Distributions
- Blog: Situating the social issues of image generation models in the model life cycle: a sociotechnical approach
- Blog: Human-AI Coevolution
- Blog: Fairness & Privacy in an Age of Generative AI
- Blog: Towards Theoretical Understandings of Self-Consuming Generative Models
- Blog: On the Dangers of Bootstrapping Generation for Continual Learning and Beyond
- Blog: Model Collapse Demystified: The Case of Regression
- Blog: Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data
- Blog: Will Large-scale Generative Models Corrupt Future Datasets?
- Blog: Public code for Model Collapse
- Blog: Voluminous yet Vacuous? Semantic Capital in an Age of Large Language Models
- Blog: Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data
- Blog: Language Generation with Replay: A Learning-Theoretic View of Model Collapse
- Blog: Large Language Models as Search Engines: Societal Challenges
- Blog: How bad is training on synthetic data? A statistical analysis of language model collapse
- Blog: Correction: AI content detection in the emerging information ecosystem: new obligations for media and tech companies
- Blog: Machine-generated text detection prevents language model collapse
- Blog: AI-generated data contamination erodes pathological variability and diagnostic reliability
- Blog: Data Augmentation using LLMs: Data Perspectives, Learning Paradigms and Challenges
- Blog: Collapse or Thrive: Perils and Promises of Synthetic Data in a Self-Generating World
- Blog: Artificial Intelligence Index Report 2024
- Blog: Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferences
- Blog: Generating on Generated: An Approach Towards Self-Evolving Diffusion Models
- Blog: Self-Improving Diffusion Models With Synthetic Data
- Blog: Art and the science of generative AI: A deeper dive
- Blog: Identifying and Mitigating the Security Risks of Generative AI
- Blog: When AI Eats Itself: On the Caveats of Data Pollution in the Era of Generative AI
- Blog: Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World
- Blog: Model Collapse in the Self-Consuming Chain of Diffusion Finetuning: A Novel Perspective from Quantitative Trait Modeling
- Blog: Removing Bias from Maximum Likelihood Estimation with Model Autophagy
- Blog: Establishing Trustworthiness: Rethinking Tasks and Model Evaluation
- Blog: A linguistic analysis of undesirable outcomes in the era of generative AI
- Blog: Rate of Model Collapse in Recursive Training
- Blog: A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops
- Blog: Machine Culture
- Blog: Universality of the π2/6 Pathway in Avoiding Model Collapse
- Blog: A Survey of Post-Training Scaling in Large Language Models
- Blog: A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops
- Blog: How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse
- Blog: AI-Generated Images Introduce Invisible Relevance Bias to Text-Image Retrieval
- Blog: Analyzing and Improving Model Collapse in Rectified Flow Models
- Blog: Identifying and Mitigating the Security Risks of Generative AI
- Blog: Understanding Hallucinations in Diffusion Models through Mode Interpolation
- Blog: We Need Improved Data Curation and Attribution in AI for Scientific Discovery
- Blog: When Models Don’t Collapse: On the Consistency of Iterative MLE
Showing 60 of 415 papers.
History & citation
Current · Permanent link
- The Data Cannibalization Paradox of Generative AI merged into this direction (68% overlap)
- Model Collapse and the Curse of Recursion merged into this direction (100% overlap)
- Neural Retrieval Bias Toward Synthetic Content merged into this direction (55% overlap)
Cite this direction
Latest version (a living page):
@misc{lacuna_recursive_training_degradation_045c2b20,
title = {recursive training degradation},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/recursive-training-degradation/rdr_045c2b203d21408da841bb5495bf4169},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}This exact version (version 2, frozen text):
@misc{lacuna_recursive_training_degradation_045c2b20_v2,
title = {recursive training degradation},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/recursive-training-degradation/rdr_045c2b203d21408da841bb5495bf4169/v/2},
note = {Version 2, frozen 2026-10-04. Accessed 2026-11-19},
}