IOP Publishing has published a new diamond open-access journal dedicated to artificial intelligence (AI). AI Horizons responds to the growing demand for research covering AI for engineering and science.
The journal is published in partnership with the Eastern Institute of Technology (EIT) in Ningbo, the Zhejiang Association for Science and Technology and the World Association of Young Scientists.
Shiyi Chen, who is president of the EIT, is editor-in-chief of AI Horizons. Chen has worked at Johns Hopkins University in the US, Peking University and the Southern University of Science and Technology. His research has focused on turbulence and computational fluid dynamics but more recently has extended to AI for engineering and industrial applications.
Here, Chen discusses the growing impact of AI in science and future plans for the journal.
How is AI changing science?
AI is changing how we work with data and how we develop scientific models. It can help us identify patterns and explore problems that are difficult to tackle with conventional methods alone. The scientific value depends on what we learn from the results and how well they hold up to testing. For me, the most promising work combines AI with physical understanding and careful comparison with experiments or observations.
Which applications of AI excite you most, and why?
AI for engineering and industrial applications interests me most, especially turbulence and complex flows. Turbulence has been central to my research for many years and it is a demanding test for AI because it involves interactions across a wide range of scales. I am interested in how AI can work with physical models to improve simulations and make them more useful in engineering design. Integrating this work into computer-aided engineering and industrial software could help engineers evaluate designs more efficiently.
What gap does AI Horizons fill, and what will make it distinctive?
AI Horizons will focus on AI that advances scientific understanding and engineering practice. Its scope brings together AI for science and engineering, spatiotemporal intelligence, embodied AI, and quantum intelligence. I would be particularly interested in a paper that uses AI to resolve a difficult scientific question or makes a previously impractical engineering calculation feasible.
We also welcome fundamental advances in AI methods. Authors should explain what their work makes possible and provide convincing evidence for their claims, including where the method reaches its limits. I want researchers from different fields to find work here that they can test and build on.
How will the journal bring together different disciplines?
A paper that combines AI with another field needs reviewers who understand both the methods and the scientific problem and we will seek that combination of expertise in peer review. We also hope to commission reviews and perspectives that help readers understand useful ideas from other fields. The aim is to make the papers accessible and valuable to researchers beyond their immediate specialty.
How important is it that the journal is diamond open access?
It matters a great deal. Readers can access the papers freely, and authors do not pay article processing charges because the journal’s sponsors cover the publication costs. This is particularly valuable for researchers with limited funding. I want the decision to submit to AI Horizons to depend on whether the journal is right for the work, without publication costs becoming an obstacle.
Physicists discuss the future of machine learning and artificial intelligence
Read more
Do you think AI will increase the need for new article formats or other publishing innovations?
Yes. For many AI studies, readers need access to the code, data, and details of model training to assess the results properly. Papers should make it easier to reproduce the work and test it in other settings. I would like us to explore formats that support this, including interactive material that helps readers examine the results. Any updates to published work would need clear versioning and appropriate editorial oversight.
AI is developing very quickly. How will the journal balance speedy publication with rigorous peer review?
Authors deserve timely decisions and useful feedback. We will aim to keep editorial handling efficient and involve suitable reviewers early. Editors also need to give reviewers enough time to assess the work and authors enough time to address substantive questions. We should avoid unnecessary delays while ensuring that the published findings have been examined carefully.
What do you hope the journal will achieve in the coming years?
I hope AI Horizons becomes a journal researchers trust for sound, original work and turn to for ideas they can use in their own fields. I would be especially pleased to see papers published here lead to collaborations between AI researchers and scientists or engineers working on difficult problems. The journal will earn its reputation through the quality of its papers and how well it serves those researchers.