Call For Competition Proposals
The IJCNN 2027 Competition Program invites proposals for challenging, accessible, reproducible, fair, and community relevant competitions within the scope of neural networks, the foundations of AI and intelligent applications. Proposals should address timely scientific, engineering, or societal problems and create benchmarks with value beyond the conference.
Submission Portal Opening Soon
Topics include, but are not limited to:
Robustness and fairness
Efficient, edge, and low-resource learning
Continual and online learning
Autonomous agents and embodied intelligence
Multimodal and graph learning
Time-series analysis
Scientific AI
Healthcare, climate, energy, communications, and critical infrastructure
Privacy-preserving, federated, and distributed learning
Emerging nature-inspired or computational-intelligence paradigms
Proposal Requirements
Proposals must be submitted as a PDF of at most 4 pages, excluding references, and should include:
1. Competition title, task, motivation, and expected impact
2. Justification for using a competition format
3. Target community, accessibility, and expected participation
4. Datasets, provenance, licenses, access, and train–test splits
5. Evaluation metrics, ranking, tie-breaking, and hidden test policy
6. External-data and pretrained-model policies
7. Baseline, starter kit, evaluation code, and platform plan
8. Anti-cheating, leakage-prevention, and fairness safeguards
9. Reproducibility and winner-verification requirements
10. Ethics, privacy, consent, copyright, and misuse considerations
11. Organizers, affiliations, contacts, and short biographies
12. Funding, sponsorship, prizes, schedule, and post competition plan.
Organizer Responsibilities
Accepted organizers must provide:
A public competition webpage and clear participation rules
Timely access to training data and an appropriate test set
Baseline code or a starter kit whenever feasible
Transparent scoring, submission limits, and ranking procedures
Verification of top-ranked solutions
Participant support, FAQ, and regular communication
A final challenge report and long-term resource archive
Important Note: Organizers, co-organizers, and their direct advisees may not compete in the competition they organize. Competitions should normally run for approximately 3–4 months. A dry-run or readiness review may be required before public launch.
Evaluation and Reproducibility
Competitions must announce primary and secondary metrics, ranking procedures, submission limits, tie-breaking rules, and any efficiency, latency, robustness, or fairness criteria. A hidden final test set is strongly recommended. Top-ranked teams may be required to provide system descriptions, external-data declarations, inference code, containers, resource statements, model cards, or other materials needed to verify results.
Review Criteria
Proposals will be evaluated according to:
Relevance to IJCNN and significance of the challenge
Dataset quality, legality, accessibility, and documentation
Rigor and transparency of the evaluation protocol
Feasibility and readiness of the organizing team
Accessibility and expected community participation
Reproducibility, ethics, fairness, and responsible-AI safeguards
Expected long-term benchmark and community value
Competition Papers
Subject to conference policy and peer review, up to three leading teams may be invited to submit papers for possible inclusion in the IJCNN 2027 proceedings on IEEE Xplore. Organizers may also submit a competition report describing the dataset, task, baselines, evaluation, participation, and results. Accepted papers must be registered and presented at the conference.