Ryze: Evidence-Enriched Data Synthesis from Biomedical Papers
On the map: Synthetic data and self-improvement for LLMs → Medical synthetic data and agents
Abstract (arXiv)
General-purpose VLMs remain unreliable for biomedical research because valid answers in scientific papers depend on evidence split across figures, tables, charts, captions, and referring text. Existing post-training pipelines are bottlenecked by costly expert annotation and by synthetic data that drops this evidence structure. We present Ryze, a fully automated system that converts raw biomedical papers into an evidence-enriched training set and a domain-specialized VLM. Ryze synthesizes QA pairs with complete supporting evidence (visual element, caption, extracted structure, and referring paragraphs), reduces layout and OCR errors via chart/table-aware extraction and LLM-based cleansing, and applies a progress-gated post-training strategy combining supervised fine-tuning with reinforcement learning. Starting from Qwen3-VL-8B, Ryze produces BioVLM-8B at under USD 200, achieving 48.0% weighted accuracy on LAB-Bench, outperforming the base model by +12.6 percentage points (pp) and surpassing GPT-5.2 by +3.8 pp. We release Ryze as open source together with the trained BioVLM-8B model.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Medical synthetic data and agents, one of 13 directions in Synthetic data and self-improvement for LLMs. Builds on 10 of its 18 references that are in the map. First posted on arXiv 30 May 2026.
Builds on
10 of 18 references.
References · 10
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
- 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
- 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,442
- 2020What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Examscited by 1,324
- 2023LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Daycited by 1,287
- 2019PubMedQA: A Dataset for Biomedical Research Question Answeringcited by 1,271
- 2020LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understandingcited by 406
- 2023Nougat: Neural Optical Understanding for Academic Documentscited by 240
- 2025DeepSeek-OCR: Contexts Optical Compressioncited by 174
- 2024LAB-Bench: Measuring Capabilities of Language Models for Biology Researchcited by 138
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Weaving Multi-Source Evidence for Biomedical Reasoning: The BioMedHop Benchmark and BioWeave Framework
- 2025Query-driven Document-level Scientific Evidence Extraction from Biomedical Studies
- 2025Aligning LLMs with Biomedical Knowledge using Balanced Fine-Tuning
- 2026When Retrieval Helps and Distracts: Evaluating Evidence-Generating LLMs for Biomedical Claim Verification
- 2025Enhancing Study-Level Inference from Clinical Trial Papers via Reinforcement Learning-Based Numeric Reasoning
- 2026Large Language Models Meet Biomedical Knowledge Graphs for Mechanistically Grounded Therapeutic Prioritization
- 2026Not Just Reason, Not Just Scan: Reinforcement Learning for Proactive Scientific Error Verification over Academic Paper
- 2026AI Co-Scientist for Knowledge Synthesis in Medical Contexts: A Proof of Concept