Academic Journal
Quartile not availableCAAI Artificial Intelligence Research
About CAAI Artificial Intelligence Research
CAAI Artificial Intelligence Research: Journal Profile
CAAI Artificial Intelligence Research, abbreviated CAAI AIR, is an open-access, peer-reviewed journal in artificial intelligence published by Tsinghua University Press (Beijing, China) and released exclusively on the publisher's SciOpen platform. It is jointly sponsored by the Chinese Association for Artificial Intelligence (CAAI) and Tsinghua University. The journal carries print ISSN 2097-194X and electronic ISSN 2097-3691 — the JournalsWorld record profiled here is the one registered under the electronic ISSN — together with the Chinese domestic publication number CN 10-1840/TP. It was launched in 2022 as one of the "high-start" new-journal projects of China's Excellence Action Plan for science and technology journals. This profile identifies the journal by its ISSNs and publisher throughout, because CAAI sponsors more than one AI periodical: it must not be confused with CAAI Transactions on Intelligence Technology, a different journal (print ISSN 2468-6557, electronic ISSN 2468-2322) published through a different publishing arrangement.
Overview and Launch
CAAI Artificial Intelligence Research published its inaugural issue in 2022. In his message in that first issue, the Editor-in-Chief introduced the journal as a new, high-starting-point journal project within the Excellence Action Plan of China Science and Technology Journals, created to reflect the state of the art in artificial intelligence and its applications at a moment when deep learning, big data and expanding computing resources had opened a new phase in the field, and when Chinese researchers had become a major force in world AI research. The inaugural issue set the pattern that the journal has kept: a compact set of original research and review articles — three of each in the first issue — selected through a double-blinded peer-review process. The journal's article numbering is visible in its DOIs, which carry the prefix 10.26599/AIR (for example 10.26599/AIR.2022.9150000 for the inaugural editorial matter), registered through Crossref. Issues have appeared for every year from 2022 onward, with volume 3 corresponding to 2024, and the journal's online archive on SciOpen runs continuously to the current year.
Aims and Scope
The journal's stated aim is to publish state-of-the-art achievements in artificial intelligence and its applications, and to give AI researchers and practitioners an international stage for sharing significant findings. Its official scope names six anchor areas: knowledge intelligence; perceptual intelligence; machine learning; behavioral intelligence; brain and cognition; and AI chips and applications. Original research articles and review articles across these areas — and on closely related topics beyond the named list — are welcomed. The content published so far illustrates the breadth of that remit: the early volumes include work on phonetic-semantic pretraining for robust speech recognition, surveys of noncooperative games and distributed Nash equilibrium seeking over multi-agent networks, reinforcement-learning approaches to combinatorial optimisation and object tracking, human–machine teaming, and interactive embodied intelligence. The journal also runs topical collections and special issues with open calls for papers — a special issue on emerging techniques for processing and analysing graph data, with a submission deadline at the end of 2025, is one recent example listed on its SciOpen pages — which authors can use to target themed readerships within the journal's wider scope.
Sponsorship, Publisher and Platform
Governance of the journal is shared. The Chinese Association for Artificial Intelligence (CAAI) — China's national learned society for AI — and Tsinghua University are the journal's sponsors, and the SciOpen journal pages present both organisations as the journal's society, partner and affiliation. Publishing is handled by Tsinghua University Press, the university's academic publishing house, and the journal is released exclusively on SciOpen, Tsinghua University Press's platform for its science and technology journals. SciOpen is itself supported by the China Science and Technology Journal Excellence Action Plan. Manuscripts are handled through the publisher's submission route linked from SciOpen (the ScholarOne Manuscripts system at mc03.manuscriptcentral.com), and the journal's editorial office and collaboration contacts are listed on the SciOpen pages, based at the Press's offices in the Xueyan Building on Shuangqing Road, Haidian District, Beijing. For readers, the practical consequence of the SciOpen arrangement is that the version of record, the article PDFs and the journal's metrics all live on sciopen.com rather than on a general commercial publisher platform.
Editorial Leadership
The journal's Editor-in-Chief is Qionghai Dai of Tsinghua University, a senior figure in computational imaging and AI, who authored the inaugural editorial message. He leads an editorial board whose full membership is published on the journal's SciOpen pages, alongside an editorial office at Tsinghua University Press that manages submissions, production and journal collaboration enquiries. The board structure follows the common pattern for Chinese society journals: the sponsoring association and university supply the academic leadership, while the press supplies professional publishing staff. Authors evaluating the journal can inspect the current board composition directly on SciOpen, since board rosters are updated there as appointments change; this profile names only the Editor-in-Chief, whose tenure is stated on every journal page on the platform.
Publication Frequency and Article Types
The journal was announced at launch as a quarterly publication, and its content is organised in annual volumes (volume 1 in 2022, volume 2 in 2023, volume 3 in 2024). At the same time, the publisher describes the journal as appearing with continuous online publication: articles are posted on SciOpen as soon as they are accepted and produced — the platform's "Just Accepted" stream — rather than being held for a completed issue, and are then collected into the year's volume. The journal publishes original research articles and review articles; those are the two article categories named in its aims and scope, and the inaugural issue's balance of three research papers and three reviews shows the weight the journal gives to authoritative survey content alongside new results. Occasional editorials and special-issue introductions appear in addition. Manuscripts must be in English, which DOAJ records as the journal's accepted language.
Peer-Review Model
CAAI Artificial Intelligence Research uses double-anonymous (double-blind) peer review, as recorded in its DOAJ listing: neither authors nor reviewers know each other's identities during assessment. The inaugural editorial confirms that the first issue's papers passed through a double-blinded process, and DOAJ additionally records that the journal checks submissions for plagiarism. DOAJ's figure for the journal's overall timeline is an average of about 16 weeks from submission to publication. The SciOpen journal page publishes the platform's own current service metrics, which at the time this profile was prepared showed averages of roughly 20 days to first decision, 68 days to acceptance and 30 days to publication; these platform figures are updated by the publisher and should be read as indicative operational statistics rather than guarantees. As with any double-blind venue, authors should prepare an anonymised manuscript file, keeping identifying acknowledgements and self-references out of the review copy in line with the journal's submission guidelines on SciOpen.
Open Access, Licensing and Fees
The journal is fully open access, and its funding model is unusually favourable to authors: both the publisher's journal pages and the DOAJ record state that there are no article processing fees — publishing in CAAI Artificial Intelligence Research is completely free of charge for authors, with the journal's costs carried by its sponsors and publisher rather than by authors or readers. Articles are published under a Creative Commons Attribution licence (CC BY), and DOAJ records that authors retain unrestricted copyright and publishing rights in their work. Articles carry DOIs as their permanent identifier. The journal has been open access since its first issue in 2022. For researchers without grant funding for APCs — and for those whose funders do not reimburse charges at commercial open-access journals — a no-fee, society-sponsored, CC BY journal backed by a major university press is a materially different proposition from the standard gold open-access model, and it is one of this title's defining characteristics.
Indexing and Abstracting
The journal is listed in the Directory of Open Access Journals (DOAJ), which verifies its open-access practices, licensing, peer-review model and fee policy as described above. The SciOpen journal site maintains a dedicated abstracting and indexing page for the title, and the journal's articles are discoverable through SciOpen itself and through the Crossref DOI infrastructure. At the time this profile was prepared, inclusion in Scopus or the Web of Science could not be verified from an official source, and no Impact Factor or CiteScore is claimed for the journal here; authors whose institutions require those specific indexes should check the Scopus source list and Clarivate's Master Journal List directly for ISSN 2097-3691 before submitting. This caution matters because the journal's better-known sibling title, CAAI Transactions on Intelligence Technology, does carry citation metrics of its own — those metrics belong to the Transactions and must not be attributed to CAAI Artificial Intelligence Research.
Submission Process
Submissions are made online through the manuscript system linked from the journal's SciOpen homepage ("Submit Manuscript"), which routes to the publisher's ScholarOne Manuscripts portal. Before submitting, authors should read the journal's submission guidelines on SciOpen, prepare an anonymised manuscript suitable for double-blind review, and ensure the work falls squarely within the six scope areas or their close neighbours. Calls for papers for topical collections are posted on the SciOpen journal pages with their own deadlines, and authors responding to a call should follow the instructions in the call. Because there are no APCs, the financial side of submission is simple: the costs to plan for are the author's own time in revision, across a process whose published averages suggest a first decision in a matter of weeks.
Readership and Fit
The journal addresses AI researchers and practitioners internationally, with a natural centre of gravity in China's large AI research community and in the networks of CAAI and Tsinghua University. Its free-to-read, free-to-publish model and its English-language policy are designed to attract submissions from worldwide, and its scope — spanning knowledge and perceptual intelligence, machine learning, behavioural intelligence, brain and cognition, and AI hardware — suits authors whose work crosses subfield boundaries or speaks to AI as a whole rather than to a single narrow specialism. For authors, the fit is strongest for substantial original research and for authoritative reviews that can serve as reference points in a fast-moving area. Authors should weigh the journal's youth — it has published only since 2022, and its citation-metric record is still developing — against its institutional backing, rigorous double-blind review and zero-fee open-access model when choosing where to submit.
Distinguishing Similar Titles
Two confusions are worth preventing explicitly. First, CAAI Transactions on Intelligence Technology (ISSN 2468-6557 / 2468-2322) is a separate, older CAAI journal published in partnership with a commercial publisher; citations, metrics and submission instructions for that title do not apply to CAAI Artificial Intelligence Research. Second, CICAI (the CAAI International Conference on Artificial Intelligence) is a conference whose proceedings appear in Springer's Lecture Notes in Computer Science; it is not this journal. The journal profiled here is identified uniquely by the combination of Tsinghua University Press as publisher, SciOpen as platform, print ISSN 2097-194X and electronic ISSN 2097-3691.
The SciOpen Platform Experience
Reading and using the journal happens entirely on SciOpen, and the platform shapes the reader's experience in a few specific ways. The journal homepage brings together the aims and scope, the editorial board and editorial office pages, the journal's publishing-ethics statements, its abstracting and indexing page, and information about its sponsoring society and publisher. Content is browsable by issue and year, through a "Just Accepted" stream for the newest papers, a latest-issue view, and topical collections with their calls for papers; alert and RSS options let readers follow new content without revisiting the site. Each article has a permanent DOI under the 10.26599 prefix used by Tsinghua University Press titles, with the journal code AIR embedded in the suffix, and articles are readable in HTML and downloadable as PDF at no charge. Journal collaboration, technical support and media contacts for the Press are published in the site footer, and the Press's offices are given as Floor 6, Tower B, Xueyan Building, Shuangqing Road, Haidian District, Beijing 100084, China. Authors should use the "For Authors" area of the same pages — submission guidelines and the manuscript-submission entry point — rather than emailing manuscripts, since the journal's workflow runs through its online system.
Position Within Chinese and International AI Publishing
CAAI Artificial Intelligence Research was created under a national programme — the Excellence Action Plan of China Science and Technology Journals — whose purpose is to build Chinese-published journals capable of standing alongside the established international titles in their fields, rather than seeing the country's best work flow exclusively to overseas publishers. Its design choices follow from that ambition: sponsorship by the national AI association and a leading university, a professional university press as publisher, an English-language policy aimed at an international authorship and readership, double-blind review, immediate open access under CC BY, and — most unusually by international commercial standards — the complete absence of author fees, funded as a matter of public and institutional investment in the journal. For authors outside China, the title offers a no-cost route into a venue with serious institutional backing and a readership anchored in one of the world's largest AI research communities; for the global field, it is one of the clearest examples of the society-and-university-press model being applied to artificial intelligence at national scale. Its long-term standing will depend on the citation record it accumulates, which readers should assess from the journal's own growing archive rather than from the metrics of any similarly named title.
What the Early Volumes Show
The journal's opening content is a useful guide to what its editors consider in scope and up to standard. The inaugural issue paired three original research papers with three reviews, among them a phonetic-semantic pretraining model aimed at robust speech recognition in noisy conditions — perceptual intelligence applied to a concrete engineering problem — and a survey of noncooperative games and distributed Nash equilibrium seeking across multi-agent networks, a behavioural-intelligence and control-theory topic. Subsequent volumes have added decision-making in team-adversary games, deep-reinforcement-learning object tracking, reinforcement learning for robotic assembly, human–machine teaming and interactive embodied intelligence, with contributions from authors inside and outside China. Two patterns stand out for prospective authors. Review and survey articles are a genuine pillar of the journal, not filler, so a well-constructed survey of a fast-moving area is a realistic submission. And the published mix spans theory (game theory, optimisation), systems-facing work (speech, robotics) and conceptual pieces (embodied intelligence), confirming that the six named scope areas operate as an open framework rather than a narrow checklist. Authors can inspect every published paper free of charge on SciOpen — the most direct way to judge fit before writing a submission.
A last practical observation concerns timing and expectations. The journal is young, its archive is still small enough to browse in full, and its published service metrics — a first decision in roughly three weeks on the SciOpen page, and a DOAJ-recorded average of about sixteen weeks from submission to publication — describe a process considerably faster than the multi-season cycles common at large commercial AI journals. Authors who need a citable, openly accessible version of record for a grant report or a thesis-by-publication can plan around those figures, while remembering that they are averages published by the platform, not contractual commitments, and that double-blind review timelines depend on reviewer availability in the relevant specialism.
Key Facts at a Glance
- Full title: CAAI Artificial Intelligence Research (CAAI AIR)
- Print ISSN: 2097-194X · Electronic ISSN: 2097-3691 · CN: 10-1840/TP
- Publisher: Tsinghua University Press, Beijing, China (exclusively on SciOpen)
- Sponsors: Chinese Association for Artificial Intelligence (CAAI) and Tsinghua University
- Publishing since: 2022; quarterly volumes with continuous online publication
- Editor-in-Chief: Qionghai Dai (Tsinghua University)
- Scope: knowledge intelligence; perceptual intelligence; machine learning; behavioral intelligence; brain and cognition; AI chips and applications
- Peer review: double-anonymous; plagiarism screening (per DOAJ)
- Fees: no article processing charges for authors
- Licence: CC BY; authors retain unrestricted copyright (per DOAJ)
- Language: English
- Indexing: DOAJ
Profile prepared from the journal's official SciOpen pages (journal homepage, aims and scope, inaugural editorial message), its DOAJ record and Tsinghua University Press information. Editorial rosters, metrics and indexing change over time; verify current details on SciOpen before submitting.
Journal Metrics
A quartile may vary by subject category. Values without a source or reporting year are unverified historical entries. Verify the current Journal Impact Factor with Clarivate or the publisher before using it.
Aims & Scope
The publisher’s official aims and scope have not yet been verified for this profile. Use the journal website to check subject fit and accepted article types before submitting.
Recent Research Articles
Latest publications matched automatically by ISSN.
Underwater small object detection: A survey
Peixian Zhuang, Yuchun Wang, Yihang Wang, Fei Liu et al.
2026-10-14 · DOI: 10.26599/air.2026.9150010Advances in EEG-based neural mechanism analysis and decoding methods for pain
Guanghao Sun, Guixing Lin, Yang Liao, Xiangcheng Li et al.
2026-10-14 · DOI: 10.26599/air.2026.9150012Anomaly detection in long-term time series with Transformer and variational autoencoder
Xiankun Shi, Yibo Li, Ziqi Wang, Junqi Zhang et al.
2026-10-14 · DOI: 10.26599/air.2026.9150011Towards Robust Deep Reinforcement Learning against Environmental State Perturbation
Chenxu Wang, Bang You, Huaping Liu
2026-10-09 · DOI: 10.26599/air.2026.9150014Deep learning in dental image analysis: A systematic review of datasets, methodologies, and emerging challenges
Zhenhuan Zhou, Jingbo Zhu, Yuchen Zhang, Xiaohang Guan et al.
2026-10-09 · DOI: 10.26599/air.2026.9150006PEMUTA: Pedagogically-Enriched Multi-Granular Undergraduate Thesis Assessment
Jialu Zhang, Qingyang Sun, Qianyi Wang, Risa Higashita et al.
2026-09-21 · DOI: 10.26599/air.2026.9150013Neural Network-Based Fixed-Time Control for Flexible Two-link Manipulators with State Constraints and Input Deadzone
Hejia Gao, Tanyu Chen, Jiangxu Liu, Dengxiang Chang et al.
2026-09-02 · DOI: 10.26599/air.2026.9150009Robust Event-Triggered MPC with Parallel Optimization and Disturbance Rejection for Mobile Robots
Hejia Gao, Yanling Guo, Puhai Yang, Changyin Sun et al.
2026-08-25 · DOI: 10.26599/air.2026.9150008Transformer-Based Hyperspectral Feature Learning Model for Early Detection of Wheat Yellow Rust
Geoffry Mutiso, Charles Munyao, John Ndia
2026-07 · DOI: 10.26599/air.2026.9150007Masked Denoising Diffusion for Accurate Anomaly Recognition and Reconstruction
Haigang Zhang, Zican Baolin, Zhitao Wu, Wei Zhang et al.
2026-07 · DOI: 10.26599/air.2026.9150005GE-LSTM-Net: A global-enhanced LSTM network with attention for spatiotemporal feature extraction in EEG-based vigilance estimation
Hao Lan, Meng Tang, Xiangyu Ju, Ming Li et al.
2026-07 · DOI: 10.26599/air.2026.9150001Curvature-Aware Model Aggregation for Decentralized Federated Learning via Parallel Transport on Discrete Manifolds
Wei Gao
2026-07 · DOI: 10.26599/air.2026.9150004FG-RAG: A fine-grained semantic optimization framework for retrieval-augmented generation
Mingan Xu, Tingjie Liu, Youhua Xia
2026-06 · DOI: 10.26599/air.2026.9150003GraphCredit: Enhancing credit explainability via gated graph reasoning and causal LLM attribution
Bocheng Wang, Di Han, Yuchang Zou, Jianqing Li et al.
2026-06 · DOI: 10.26599/air.2026.9150002Image field-of-view enhancement: An overview
Peixian Zhuang, Chenyu Wang, Fei Liu, Xinheng Zhang et al.
2026-06 · DOI: 10.26599/fie.2026.9270080Silhouette-driven contrastive learning with hierarchical neighborhood structure for unsupervised clothes change person re-identification
Mingkun Li, Peng Xu
2026-03 · DOI: 10.26599/air.2025.9150050A survey on deep learning techniques for image and video feature aggregation
Md Majedul Islam, Sai Prakash Reddy Rao, Selena He
2026-01 · DOI: 10.26599/air.2025.9150054Evaluation study on SAM 2 for class-agnostic instance-level segmentation
Jialun Pei, Zhangjun Zhou, Tiantian Zhang
2026-01 · DOI: 10.26599/air.2025.9150055Optimal strategies of blockchain-enabled platform supply chain with manufacturer’s altruistic preference
Guoyu Bi, Chunming Xu, Chenchen Wu
2026-01 · DOI: 10.26599/air.2025.9150052Underwater light field imaging: A survey
Peixian Zhuang, Xinheng Zhang, Junjie Tong, Chenyu Wang et al.
2025-12 · DOI: 10.26599/air.2025.9150051Reviews
Community Reviews
Version History
October 14, 2026 at 3:02 am
October 14, 2026