Self-Improvement in Multimodal Large Language Models: a survey.

Deng, S., Wang, K., et al. (2025, October 3).

arXiv.org.

https://arxiv.org/abs/2510.02665

Abstract

Recent advancements in self-improvement for Large Language Models (LLMs) have efficiently enhanced model capabilities without significantly increasing costs, particularly in terms of human effort. While this area is still relatively young, its extension to the multimodal domain holds immense potential for leveraging diverse data sources and developing more general self-improving models. This survey is the first to provide a comprehensive overview of self-improvement in Multimodal LLMs (MLLMs). We provide a structured overview of the current literature and discuss methods from three perspectives: 1) data collection, 2) data organization, and 3) model optimization, to facilitate the further development of self-improvement in MLLMs. We also include commonly used evaluations and downstream applications. Finally, we conclude by outlining open challenges and future research directions.

Here are some thoughts that summarize this paper. MLLMs are learning to improve without human oversight.

This survey presents the first comprehensive overview of self-improvement in Multimodal Large Language Models (MLLMs), a rapidly emerging paradigm that enables models to autonomously generate, curate, and learn from their own multimodal data to enhance performance without heavy reliance on human annotation. The authors structure the self-improvement pipeline into three core stages: data collection (e.g., via random sampling, guided generation, or negative sample synthesis), data organization (including verification through rules, external or self-based evaluators, and dataset refinement), and model optimization (using techniques like supervised fine-tuning, reinforcement learning, or Direct Preference Optimization). The paper reviews representative methods, benchmarks, and real-world applications in domains such as math reasoning, healthcare, and embodied AI, while also outlining key challenges—including modality alignment, hallucination, limited seed model capabilities, verification reliability, and scalability. The goal is to establish a clear taxonomy and roadmap to guide future research toward more autonomous, general, and robust self-improving MLLMs.

AI models fed AI-generated data quickly spew nonsense

Elizabeth Gibney

nature.com

News: July 24, 24

Training artificial intelligence (AI) models on AI-generated text quickly leads to the models churning out nonsense, a study has found. This cannibalistic phenomenon, termed model collapse, could halt the improvement of large language models (LLMs) as they run out of human-derived training data and as increasing amounts of AI-generated text pervade the Internet.

“The message is, we have to be very careful about what ends up in our training data,” says co-author Zakhar Shumaylov, an AI researcher at the University of Cambridge, UK. Otherwise, “things will always, provably, go wrong”. he says.” The team used a mathematical analysis to show that the problem of model collapse is likely to be universal, affecting all sizes of language model that use uncurated data, as well as simple image generators and other types of AI.

The researchers began by using an LLM to create Wikipedia-like entries, then trained new iterations of the model on text produced by its predecessor. As the AI-generated information — known as synthetic data — polluted the training set, the model’s outputs became gibberish. The ninth iteration of the model completed a Wikipedia-style article about English church towers with a treatise on the many colours of jackrabbit tails (see ‘AI gibberish’).

The info is here.

Here are some thoughts:

This article highlights a concerning phenomenon known as model collapse, which occurs when artificial intelligence (AI) models are trained on text generated by other AI models. This recursive training leads to a degradation in the quality of outputs, ultimately resulting in nonsensical responses. Researchers demonstrated that as AI-generated content increasingly permeates the internet, the reliance on this synthetic data could stifle the advancement of large language models (LLMs) due to a lack of high-quality human-derived training data. The study revealed that even before complete collapse, models trained on AI-generated texts tend to forget less frequent information, which poses significant risks for fair representation of marginalized groups.

Said differently: AI garbage in, AI garbage out.