aiqbench analysis
OpticalDNA reframes genomic modeling as an OCR-style document understanding task, addressing the inefficiency of treating DNA as a simple linear sequence. By converting genomic data into structured visual layouts, the framework achieves superior performance on sequences up to 450k bases while using nearly 20 times fewer effective tokens than current baselines. This approach demonstrates that vision-language architectures can significantly reduce computational overhead and improve compression for long-context biological data.
Rethinking Genomic Modeling Through Optical Character Recognition
arXiv:2602.02014v2 Announce Type: replace-cross Abstract: Recent genomic foundation models largely adopt large language model architectures that treat DNA as a one-dimensional token sequence. However, exhaustive sequential reading is structurally misaligned with sparse and discontinuous genomic semantics, leading to wasted computation on low-information background and preventing understanding-driven compression for long contexts. Here, we present OpticalDNA, a vision-based framework that reframes genomic modeling as Optical Character Recognition (OCR)-style document understanding. OpticalDNA renders DNA into structured visual layouts and trains an OCR-capable vision--language model with a visual DNA encoder and a document decoder, where the encoder produces compact, reconstructible visual tokens for high-fidelity compression. Building on this representation, OpticalDNA defines prompt-conditioned objectives over core genomic primitives-reading, region grounding, subsequence retrieval, and masked span completion-thereby learning layout-aware DNA representations that retain fine-grained genomic information under a reduced effective token budget. Across diverse genomic benchmarks, OpticalDNA consistently outperforms recent baselines; on sequences up to 450k bases, it achieves the best overall performance with nearly 20$\times$ fewer effective tokens, and surpasses models with up to 985$\times$ more activated parameters while tuning only 256k trainable parameters.
Abstract:Recent genomic foundation models largely adopt large language model architectures that treat DNA as a one-dimensional token sequence. However, exhaustive sequential reading is structurally misaligned with sparse and discontinuous genomic semantics, leading to wasted computation on low-information background and preventing understanding-driven compression for long contexts. Here, we present OpticalDNA, a vision-based framework that reframes genomic modeling as Optical Character Recognition (OCR)-style document understanding. OpticalDNA renders DNA into structured visual layouts and trains an OCR-capable vision--language model with a visual DNA encoder and a document decoder, where the encoder produces compact, reconstructible visual tokens for high-fidelity compression. Building on this representation, OpticalDNA defines prompt-conditioned objectives over core genomic primitives-reading, region grounding, subsequence retrieval, and masked span completion-thereby learning layout-aware DNA representations that retain fine-grained genomic information under a reduced effective token budget. Across diverse genomic benchmarks, OpticalDNA consistently outperforms recent baselines; on sequences up to 450k bases, it achieves the best overall performance with nearly 20$\times$ fewer effective tokens, and surpasses models with up to 985$\times$ more activated parameters while tuning only 256k trainable parameters.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
arXiv:2602.02014 [cs.CV]
(or arXiv:2602.02014v2 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2602.02014 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 279 words
Related coverage
What is document AI?
Databricks · Jun 15, 2026
Systematic Literature Review of Machine Learning Models and Applications for Text Recognition
arXiv · Aug 28, 2026
Reading or Guessing? Visual Grounding Failures of Vision-Language Models for OCR in Ancient Greek Editions
arXiv · Sep 9, 2026
From Pixels to Structure: Lightweight Vision-Language Models for Document OCR and Structured JSON Extraction
arXiv · Oct 10, 2026