Data-Centric Artificial Intelligence: Models and Applications

Dear Colleagues,

In recent years, the paradigm of Artificial Intelligence (AI) has undergone a significant shift from model-centric approaches that prioritize architectural innovation to data-centric AI, which emphasizes the quality, structure, curation, and utilization of data as the cornerstone of intelligent systems. As AI systems are increasingly deployed in real-world applications, it is evident that high-quality, well-structured, and contextually relevant data often plays a more decisive role in system performance than model complexity alone.

Data-centric AI encompasses a broad spectrum of research directions, including data labeling and annotation strategies, data preprocessing and augmentation, bias and fairness mitigation, data versioning, dataset engineering, and the development of evaluation frameworks centered around data quality. Crucially, these directions are deeply rooted in mathematical disciplines such as optimization theory, statistical inference, information theory, probability theory, and graph theory. Data-centric AI also intersects with emerging concerns regarding the privacy, transparency, and sustainability of AI systems. The growing availability of heterogeneous, multimodal, and large-scale datasets, from user-generated content and sensor networks to scientific and enterprise data, further underscores the need for principled, scalable, and application-aware data-centric methodologies grounded in rigorous mathematical frameworks.

This Special Issue aims to bring together cutting-edge research that advances the theory, methodology, and real-world deployment of data-centric Artificial Intelligence, with a particular emphasis on the underlying mathematical principles and novel algorithms. We invite original contributions that explore novel models, algorithms, systems, and applications where data is treated as a primary design and optimization target.

Topics of interest include, but are not limited to, the following:

1. Mathematical foundations and theories of data-centric AI, including algorithms and methods for data quality, data utility, and data value assessment;
2. Data curation, cleaning, and augmentation techniques tailored for AI training and evaluation;
3. Label-efficient learning paradigms, such as weak supervision, self-supervision, and active learning in data-centric contexts;
4. Bias, fairness, and robustness through data-centric interventions modeled and mitigated using statistical, geometric, or causal inference methods;
5. Data versioning, lineage, and reproducibility in AI development pipelines;
6. Multimodal and cross-domain data integration for enhanced model generalization;
7. Privacy-preserving and ethical data practices based on differential privacy and mathematical cryptography, including synthetic data generation and federated data strategies;
8. Benchmarking and evaluation methodologies centered on data characteristics rather than model architectures.

Prof. Dr. Hongzhi Wang
Dr. Chen Ye
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Mathematics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• data-centric artificial intelligence
• statistical inference problems
• statistical learning theory
• probabilistic models
• weakly supervised learning
• self-supervised learning
• federated learning
• explainable AI
• active learning
• manifold learning
• information-theoretic approaches
• optimization theory and algorithms

• Big Data Mining and Analytics with Applications in Mathematics (11 articles)