📊 There's a hidden treasure buried in scientific papers, and most of it is locked inside images. Experimental data embedded in charts and graphs can be read by human eyes, but processing it at scale with computers has been difficult. A team from Tohoku University and the University of Tokyo has now built an AI system called DIVE, in which multiple agents work together to read scientific figures. It pulled more than 30,000 data points out of over 4,000 papers, then proposed new hydrogen storage materials in about two minutes.

The Problem: Valuable Data Trapped Inside Paper Figures

In modern materials science, data-driven AI has become a powerful tool for discovering new materials. But one bottleneck kept getting in the way: much of the experimental data researchers have accumulated over decades exists as images, namely the graphs, charts, and tables embedded in their papers.

A human researcher can look at a figure and read the data without much trouble. Getting computers to do it accurately and at scale is another matter. Existing multimodal large language models have struggled to pull data out of scientific figures with sufficient accuracy, and they handle only a limited range of figure types.

The question that drove this research was straightforward: can we free the data that's been locked inside these figures?

DIVE: An AI "Reading Team" That Works in Stages

In February 2026, a research team led by Professor Hao Li and Director Shin-ichi Orimo at Tohoku University's Advanced Institute for Materials Research (WPI-AIMR), along with Assistant Professor Ryuhei Sato at the University of Tokyo's Graduate School of Engineering, unveiled their solution: a multi-agent AI workflow called DIVE (Descriptive Interpretation of Visual Expression).

What makes DIVE different from conventional approaches is its team-based architecture. Instead of relying on a single AI to do everything at once, DIVE assigns different roles to multiple AI agents, each handling a specific step in the extraction process.

One agent focuses on visually understanding the content of a figure. Another interprets the caption, the text accompanying the figure, to grasp the scientific context. Yet another agent verifies that the extracted numerical values are scientifically consistent.

That staged approach is where the gap with the conventional single-pass method opens up.

10–30% More Accurate Than Existing Methods

On a hydrogen storage materials benchmark, DIVE recorded extraction accuracy 10–15% higher than standard multimodal AI models. Against open-source models, the improvement exceeded 30%.

Accuracy is only half of it. The team reports that the range of figure types DIVE can handle also widened substantially.

Mining 30,000 Data Points from 4,000 Papers

Using DIVE, the team processed over 4,000 scientific papers related to hydrogen storage materials. The result: more than 30,000 experimental data points, organized into a machine-readable database.

The database, called DigHyd (Digital Hydrogen Platform), is publicly available at www.dighyd.org, so researchers anywhere can pull from it for their own hydrogen storage work.

For those unfamiliar with the field, hydrogen storage materials are substances, often metals or metal alloys, that can safely absorb hydrogen at high density and release it when needed. They are considered essential technology for building a hydrogen-based clean energy infrastructure.

New Material Candidates in Just 2 Minutes

DIVE's value goes well beyond data extraction. The team also built an "inverse design" workflow on top of the DigHyd database. Inverse design flips the traditional research process: instead of testing random materials and hoping for good results, researchers specify the performance they want, and the AI works backward to find materials that match those requirements.

Here's how it works: a researcher inputs the desired material properties; the AI generates candidate compositions from the literature database; a machine learning model predicts hydrogen weight density for each candidate; and the system iteratively refines its suggestions to meet the target specifications.

From entering the requirements to receiving candidate materials, the whole loop takes about two minutes. A human doing the same work, surveying thousands of papers and narrowing down matching materials, would need weeks or months.

Beyond Hydrogen: A Platform for All Materials Research

The DIVE technology isn't limited to hydrogen storage. According to the research team, the same approach can be applied to batteries, catalysts, thermoelectric materials, and other fields where valuable experimental data sits locked in paper figures.

Future plans include expanding DIVE to handle even more diverse figure formats and developing more autonomous AI-driven material design workflows. If scientific data trapped in images can be systematically unlocked across all fields, it could mark a new era in materials research.

The work was supported by the Japan Science and Technology Agency (JST) under the hydrogen area of its GteX program, and was published in Chemical Science on February 3, 2026, local time.

What about your country?

How is AI being applied to scientific research where you live? We'd love to hear your perspective.

Update: On June 11, 2026, the team published a follow-up that draws on DigHyd data and uses symbolic regression to extract the descriptors that matter most for hydrogen storage, feeding them into material design aimed at combining high capacity with practical operating pressure. The work has moved from building the database to using it.

References