Aims and Scope

Aims

The Journal of Advanced Intelligent Computing and Informatics (JAICI) publishes research on how intelligent systems learn, reason and explain themselves, and on how they are put to work on real information problems. It takes original research, systematic and methodological reviews, and short communications.

Papers are judged on originality, sound method, reproducibility and honest evaluation against strong baselines. Theoretical, algorithmic and applied work are all welcome, provided the contribution to computing or informatics is clear.

Scope

The scope covers four subject areas. The lists below give examples and are not exhaustive; a paper outside them but inside one of the four areas will be considered.

Learning systems and architectures

  • Deep learning and neural network architectures
  • Large language models and generative AI
  • Self-supervised, contrastive and representation learning
  • Graph neural networks and multi-agent learning
  • Distributed and federated learning

Trustworthy and explainable AI

  • Explainability, causal models and counterfactual explanation
  • Robustness, calibration and uncertainty estimation
  • Fairness, privacy and security of AI systems
  • Detection of AI-generated and manipulated content

Computing paradigms and infrastructure

  • Quantum and quantum-inspired computing
  • Edge computing and IoT intelligence
  • Efficient and lightweight machine learning
  • Knowledge representation and reasoning

Informatics applications

  • Computer vision and multimodal systems
  • Medical and health informatics, including imaging and physiological signals
  • Cybersecurity analytics and threat detection
  • Text, web and social media analytics
  • Intelligent inspection and monitoring of infrastructure

Examples of papers in scope

  • Causal disentanglement with generative models for explainable chest X-ray diagnosis
  • A dynamic graph neural network that fuses beliefs across agents in collaborative diagnosis
  • Manifold-constrained data augmentation for self-supervised ECG representation learning
  • An explainable transfer learning framework for classifying concrete surface cracks
  • Classification of phishing websites from URL and domain features
  • Lightweight detection of AI-generated product reviews using stylometric features

Papers usually declined at the desk

Every submission is checked for fit, method and originality before it goes to reviewers. The following are normally declined at that stage:

  • Routine application of an existing model to a new dataset with no methodological or practical contribution
  • Results reported without baselines, a held-out test set or any measure of variance
  • Product descriptions, system manuals or feasibility notes without a research contribution
  • Claims about deployed systems that the paper gives no means to check
  • Literature summaries that offer no methodological synthesis

If you are unsure whether a paper fits, send a short pre-submission enquiry through the Contact page.

Article types

  • Original Research Article. A full-length theoretical, algorithmic, experimental or applied study with new findings.
  • Systematic or Methodological Review. A structured synthesis following PRISMA, bibliometric or a comparable method.
  • Short Communication. A concise report of a focused finding or a brief methodological note.
  • Editorial. By invitation only, written by the editors or invited contributors.

Length and format for each type are given in the Author Guidelines.