About the journal

A distinct home for intelligence and decision research.

Foundations and evaluation of intelligence and decision making. An international community with a clearly defined editorial scope.

About the journal

The Australian Journal of Artificial Intelligence and Decision Sciences provides a forum for research on learning, reasoning, preference, judgement and choice. Its community includes researchers in artificial intelligence, decision analysis, optimisation, behavioural science and related disciplines. Contributions may advance artificial intelligence, decision sciences or their integration; combining both fields is not a condition of eligibility.

The journal values explanations of why an approach works, where it fails and what its findings allow others to conclude. It welcomes mathematical and conceptual contributions alongside computational, experimental and behavioural studies. Evidence is assessed in relation to the question and claims: a theoretical paper need not add an artificial application, and an application paper should establish a contribution beyond running familiar tools on another dataset.

Aims & Scope

The Australian Journal of Artificial Intelligence and Decision Sciences advances the foundations, methods and evaluation of artificial intelligence and decision making. It welcomes research on how systems learn and reason, how preferences and decisions are structured, and how the quality and consequences of these processes can be assessed. Contributions may address AI, decision sciences, or their integration.

The journal considers mathematical, computational, experimental and behavioural research. Its interests include learning under uncertainty, optimisation, multicriteria analysis, human-AI collaboration and reliable evaluation. Application studies should produce transferable methodological knowledge or substantive insight into decision behaviour.

Submissions should explain their contribution and support claims through appropriate analysis. Empirical comparisons should justify datasets, baselines and metrics; preference-based models should examine assumptions and sensitivity where relevant. Studies primarily advancing clinical practice, business outcomes or robotic capabilities are better directed to the corresponding specialist journal.

Topics

  • Machine learning, deep learning, statistical learning and generalisation.
  • Symbolic reasoning, neuro-symbolic AI and knowledge-based inference.
  • Generative AI, language models, reinforcement learning, agents and planning.
  • Natural language processing, computer vision and multimodal learning.
  • Decision theory, behavioural decision research and decision support.
  • Multicriteria analysis, weighting, ranking, selection and prioritisation.
  • Mathematical optimisation, multiobjective methods and operational decisions.
  • Bayesian inference, fuzzy modelling and uncertainty quantification.
  • Preference elicitation, group decisions, consensus and human-AI collaboration.
  • Explainability, fairness, accountability and governance of intelligent decisions.
  • Benchmarking, evaluation design, robustness, sensitivity and reproducibility.

Scope boundaries and manuscript fit

The central contribution should concern intelligence or decision methodology. Examples include a learning principle, a reasoning mechanism, a decision model, a preference-elicitation procedure or a defensible evaluation protocol. Domain applications are appropriate when they produce transferable methodological knowledge or substantive insight into decision behaviour.

A study primarily demonstrating patient benefit, organisational transformation, information-service quality or a physical robotic capability should normally be considered by the corresponding specialist journal. Manuscript routing depends on the principal contribution and readership, rather than the presence of AI, a ranking table or a particular dataset. Editors may explain an alternative fit; transfer of files requires the authors' agreement.

Types of Articles

Each scholarly article type follows the double-anonymous review model with at least two independent external expert reports.

01

Original Research Articles

Theoretical, methodological, computational or behavioural advances in AI or decision sciences.

State the question, contribution and relationship to established work. Describe the mathematical, computational, experimental or behavioural design sufficiently for critical assessment. Match the analysis to the claims, distinguish exploratory from confirmatory work and explain uncertainty and limitations. Formal results require precise assumptions and arguments; empirical conclusions require appropriate comparisons and evidence.

02

Benchmark and Evaluation Articles

Defensible comparisons that establish tasks, protocols or substantive evidence about methods.

Define the target task, intended users, eligible methods, dataset selection, metrics and comparison protocol. Explain tuning effort, computational resources and restrictions affecting fairness. Publish informative method characteristics and uncertainty alongside aggregate scores. A benchmark should establish useful knowledge about performance or evaluation, rather than merely produce another ordered list.

03

Data and Software Resource Articles

Documented datasets, evaluation suites or research software with demonstrated scientific value.

Identify the research need and explain how the resource addresses it. Document provenance, collection or construction, versions, quality checks, access conditions, licences and known limitations. Supply usable documentation and a substantive demonstration. Explain maintenance responsibilities and restrictions without promising permanent support that has not been arranged.

04

Replication and Robustness Studies

Tests of published claims, including informative null results and unsuccessful reproductions.

Identify the precise published claim being examined and distinguish direct reproduction, independent replication and extension. Record deviations from the original design, unavailable materials and attempts to resolve uncertainty. Report confirmatory and contradictory findings with equivalent care. A failure to reproduce should be analysed before being attributed to error or misconduct.

05

Reviews and Evidence Syntheses

Critical surveys or systematic reviews answering a defined research question.

Define a substantive question and explain how evidence was located, selected, assessed and synthesised. Systematic reviews should provide reproducible search and eligibility procedures; critical surveys should justify their coverage. Separate empirical evidence, theoretical arguments and editorial interpretation. Produce an analytical contribution beyond a catalogue of publications.

06

Brief Research Communications

Complete, focused results, diagnostic experiments or methodological corrections.

Present a complete focused contribution, such as a concise theorem, diagnostic experiment or consequential methodological correction. Brevity does not reduce the need for adequate methods, evidence and limitations. Material necessary to evaluate the central claim must remain accessible with the submission.

07

Perspectives and Research Agendas

Evidence-grounded arguments about conceptual, methodological or governance questions.

Develop an evidence-grounded argument about an emerging conceptual, methodological or governance question. Identify assumptions and competing interpretations, and distinguish supported observations from proposed directions. A research agenda should explain which questions matter, what would count as progress and what evidence is presently missing.

Ready to prepare?

The author guide explains manuscript structure, study reporting and declarations.

Read the author instructions