Editor’s Note
As Editor‑in‑Chief, my mandate for the Journal of Artificial Intelligence and Robotics (JAIR) is simple: publish research that moves intelligent machines from controlled demos to dependable systems. Scope clarity helps authors position their work, helps reviewers evaluate rigor, and helps practitioners find results they can reproduce. This page explains the topical territory we cover and the standards we expect for submissions. Throughout, you will notice a balance between algorithms and embodiment, which is where robotics research creates real value.
What We Publish in Artificial Intelligence Journal
- Original Research Articles that introduce verifiable advances in perception, learning, planning, control, human‑robot interaction, or integrated systems. Claims must be supported by transparent methods and strong baselines.
- Systematic, Scoping, or Tutorial Reviews that synthesize fragmented evidence, compare families of methods, identify gaps, and propose credible research roadmaps.
- Short Communications that rapidly share impactful findings, negative results with actionable lessons, or data and benchmark releases with complete documentation.
- Method and Benchmark Papers that provide open implementations, clear protocols, and evaluation scripts so that others can reproduce results and extend the work.
- Application Studies that validate algorithms under realistic constraints—latency, compute budgets, sensor limitations, safety requirements, and user‑in‑the‑loop dynamics.
Core Domains of Interest
We encourage submissions that connect multiple layers of the autonomy stack. The list below is indicative rather than restrictive.
- Perception and Robot Vision — 2D/3D understanding, multimodal fusion (vision, lidar, radar, audio, tactile), visual odometry, mapping, loop‑closure, and scene‑graph reasoning.
- Learning and Decision‑Making — supervised, self‑supervised, and deep learning for robotics; reinforcement and robot learning with safety constraints; imitation and offline RL; foundation models adapted to embodied tasks.
- Planning, Control, and Manipulation — motion and path planning in dynamic environments; optimal and predictive control; compliant and dexterous manipulation; grasp planning; whole‑body control; hardware‑aware and real‑time implementations.
- Human–Robot Interaction and Trust — shared autonomy, intent inference, explainable and trustworthy Artificial Intelligence, safety‑case development, and user studies with validated instruments.
- Platforms and Applications — mobile and industrial robotics, collaborative and humanoid systems, soft and swarm robotics, autonomous vehicles and aerial platforms, medical and field robotics.
Characterization and Evaluation Standards
JAIR promotes comparable, reproducible evidence. Submissions should address the following:
- Datasets and Splits — name each dataset, provide licenses, describe collection and curation, list train/validation/test splits, and explain any filtering decisions. New datasets should include detailed documentation and usage examples.
- Metrics and Protocols — justify metric choices; report multiple seeds with confidence intervals; avoid single‑scenario cherry picking. For perception tasks use mAP/AUC/top‑k as appropriate; for navigation/manipulation use success rate, time‑to‑goal, collision/contact metrics, energy, and safety violations.
- Baselines and Ablations — benchmark against strong recent methods; ablate components to show necessity; provide simple baselines to calibrate the difficulty of the task.
- Hardware and Systems — specify robot platform, sensors, calibration, communication, and compute. State inference latency, memory footprint, and power usage; describe fallback behaviors and safe‑stop conditions.
- Generalization and Robustness — evaluate cross‑domain performance, deal with sensor failure and weather/lighting, stress test rare but critical events, and analyze failure modes with visuals.
- Openness — share code and models where possible; consider containers for exact reproduction; provide video demonstrations with consistent camera geometry, overlays, and time stamps.
Editorial Priorities
- Rigor before novelty — we prize careful comparisons and transparent limitations over fragile leaderboards.
- Methods clarity — clear diagrams, parameter tables, and training details so peers can implement your approach.
- Responsible Artificial Intelligence — disclose data provenance, licensing, consents, and bias‑mitigation steps; surface safety and dual‑use risks.
- Real‑world constraints — discuss compute cost, sample efficiency, hardware availability, and maintenance.
- Impact through reuse — datasets, simulation assets, or libraries that other groups can adopt quickly.
Out of Scope
- Speculative claims without theory or evidence.
- Work that withholds essential implementation details, code, or data needed to verify the main claims.
- Advertisements for products, benchmarks designed to flatter a specific method, or purely incremental parameter sweeps.
Positioning Your Manuscript for JAIR
Use the checklist below while writing. It doubles as a talking point for cover letters.
- Problem clarity — What decision will the robot make better as a result of your contribution? In which context?
- Method transparency — Can another group reproduce your figures within ±5% using your code and instructions?
- Baseline strength — Did you compare to the most competitive recent methods and provide simple baselines?
- Robustness — What breaks the system? How often? What mitigations did you attempt?
- Ethics and safety — What are the human‑subject, privacy, or dual‑use concerns, and how did you address them?
- Openness — What will peers be able to download today to verify your claims?
Where to Go Next
Consult the Author Guidelines for template, length, and figure requirements. If your manuscript spans multiple domains, add a short ‘System Overview’ figure and a ‘Reproducibility Checklist’ appendix. For special issues, contact the editorial office with a 200‑word synopsis and a list of three suggested reviewers.
For more such blogs, please visit our Blogsite.