About Juan Li · AI4Science
Juan Li · AI4Science is an open, bilingual, source-grounded automated index of AI for Science, built for researchers and agents. It tracks how artificial intelligence enters scientific research: methods and tools, papers and experiments, researchers and institutions, and the evaluations, data, and infrastructure that shape scientific workflows.
The same public content can be read on the website, followed through RSS, connected to other tools through the public JSON API, or queried through the Agent Skill by time, field, or keyword, with source-linked bilingual summaries.
Why track the whole of AI for Science
The feed surfaces recent developments; the field, people, and organization indexes follow a research direction over time. Event pages link back to the original papers, projects, code, or data where the source provides them. Daily briefs offer a quick recap; permanent event links are useful for saving and sharing. An empty search means this index has no matching entry, not that the research does not exist.
The index treats AI as a method, and a method is only as good as its validation. It therefore looks beyond individual models and tools to the conditions that make them scientifically useful: how a research question becomes a task in which machines can participate, what data and evaluations make a result checkable, how people and AI divide the work, and what testing a result needs before it enters scientific evidence or decisions.
Many advances move faster in mathematics, molecular biology, materials science, climate science, and automated experimentation. Following the whole field shows which approaches may transfer to other disciplines, including ecology and conservation biology.
Scope and selection
We include public developments in which AI is directly used to study a scientific object, changes a scientific method, or provides an important evaluation or infrastructure for scientific discovery.
An item usually needs to meet the following conditions:
- It has a direct connection to AI for Science, not just overlapping terminology.
- Enough public evidence is available to support a checkable guide.
- It reports a new scientific capability, method, empirical result, or consequential limitation.
- Readers can follow it to a paper, preprint, project page, code or data release, talk, or original institutional announcement.
General AI product news, recruitment, computing partnerships, and institutional announcements without a concrete scientific object or result are usually excluded. A famous person, journal, or institution does not qualify an item by itself.
People and organization pages are event-based indexes. They are not complete biographies, rankings, or endorsements.
Sources and production
We prefer papers and preprints, official project and laboratory pages, code and data releases, and first-hand public material from researchers and institutions. Newsletters, aggregators, and search services may help us discover a lead, but they do not replace a checkable original source.
This is a rule-governed, AI-assisted publishing workflow. Automated tools and language models help with discovery, deduplication, classification, and bilingual drafting. Versioned gates check source identity, relevance to AI for Science, evidence sufficiency, bilingual completeness, and publication boundaries.
"Source-grounded" is not the same as "human-verified." Unless a record is explicitly marked as human-reviewed, site content is AI-generated and not individually reviewed; the two labels are never mixed.
A site summary is this project's guide, not a statement by the paper's authors. Identifiers such as a DOI appear only when they can be verified; the original source controls its license, quotation limits, and reuse conditions.
The editorial rules are defined in a versioned pipeline; not every update receives individual manual approval. The on-site guide helps readers decide what deserves closer attention. It does not replace the original paper, data, code, or official release. Readers should verify scientific conclusions against the original source before citing them.
How to use it
Juan Li · AI4Science provides the same public content through four interfaces:
- The website is best for browsing updates, daily briefs, and the histories of people and organizations.
- RSS has separate daily-brief and complete-update feeds in Chinese and English.
- The Agent Skill supports natural-language queries by time, field, and keyword, returning source-linked summaries.
- The JSON API provides a stable bilingual event structure for developers and other tools.
The Agent Skill requires no API key. After installing it, start a new session and ask any agent that supports Skills to query the public content.
AI4Science · Agent access · Agent Skill source · English daily RSS · English complete RSS · JSON API · Source coverage
Corrections and withdrawals
- Correction: factual, translation, identity, and link errors are corrected at the original URL.
- Withdrawal: if a record should no longer remain public, its URL is retained as a withdrawal page so that old links do not silently point somewhere else.
- Report a problem: use the feedback form in the page footer or email [email protected].
Copyright and reuse
Copyright in the original material remains with its authors and publishers. Juan Li · AI4Science publishes its own guides, links, indexes, and daily arrangement. It does not republish the full original work.
For short quotations, credit Juan Li · AI4Science and link to the permanent page. Please contact us before republishing guides in bulk or using the public API in another product. The website, RSS feeds, Agent Skill, and JSON API provide access to the same public content.
About Juan Li
Juan Li is a conservation biologist and spatial ecologist and an Assistant Professor at the Centre of Excellence for One Health and Environmental Sustainability, Academy of Life and Natural Sciences, Xi'an Jiaotong-Liverpool University. Her research reconstructs how the expanding human footprint reshapes where wildlife can live, using global open data, Earth observation and carefully validated AI methods.
Research · Teaching · Publications · ORCID · Google Scholar
Email: [email protected]