{"slug": "training-foundation-models-with-a-single-algorithm", "title": "Training Foundation Models with a Single Algorithm", "summary": "Researchers report that the staged pipeline used to train foundation models — pretraining, midtraining, supervised fine-tuning (SFT) and reinforcement learning (RL) — is not algorithmically fundamental, and that suitable RL objectives can be derived as maximum-likelihood objectives so a single algorithm covers the whole process. The work argues the separation between these stages, which also involves changes in training data, can be collapsed into one method.", "body_md": "Foundation models are typically trained in stages: pretraining, midtraining, supervised fine-tuning (SFT), and reinforcement learning (RL), with changes in both data and learning algorithms. We show that this algorithmic separation is not fundamental. By deriving suitable RL objectives as maximum-li", "url": "https://wpnews.pro/news/training-foundation-models-with-a-single-algorithm", "canonical_source": "https://aiflash.com/news/142544/", "published_at": "2026-10-30 15:30:09+00:00", "updated_at": "2026-10-30 15:49:30.108559+00:00", "lang": "en", "topics": ["machine-learning", "ai-research", "large-language-models", "artificial-intelligence"], "entities": [], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/training-foundation-models-with-a-single-algorithm", "markdown": "https://wpnews.pro/news/training-foundation-models-with-a-single-algorithm.md", "text": "https://wpnews.pro/news/training-foundation-models-with-a-single-algorithm.txt", "jsonld": "https://wpnews.pro/news/training-foundation-models-with-a-single-algorithm.jsonld"}}