Artificial intelligence research moves fast. The venue where you publish shapes not just citation counts but which communities read your work, who invites you to collaborate, and which companies recruit from your papers. This is your field-tested guide to the AI and ML conference landscape in 2026.
Tier 1: The Flagship Venues
These four conferences define the international AI research agenda. A first-author paper at any of them is a significant career milestone.
NeurIPS (Neural Information Processing Systems)
The largest AI conference by attendance (typically 15,000+ in-person, tens of thousands virtually). Covers deep learning, probabilistic models, reinforcement learning, neuroscience-inspired AI, and increasingly, AI ethics and social impact. Submission typically opens in May; main conference in December. Acceptance rate: ~25% across all tracks, with spotlights (<5%) and orals (<2%) for exceptional work.
ICML (International Conference on Machine Learning)
The theoretical backbone of ML research. Strong preference for mathematical rigor alongside empirical results. If your paper has clean theorems or tight proofs alongside experimental validation, ICML is an excellent target. Deadline typically in January; conference in July. Acceptance rate: ~20–25%.
ICLR (International Conference on Learning Representations)
The most community-driven top venue. Open review via OpenReview means reviews, author responses, and discussion are publicly visible during the review period. This transparency cuts both ways — strong papers benefit from public discussion, but weak papers cannot hide. Deadline typically in October; conference in May. Acceptance rate: ~30% overall (1–2% as oral presentations).
AAAI (AAAI Conference on Artificial Intelligence)
Broader AI scope than the three above — knowledge representation, automated planning, multi-agent systems, AI ethics, and applied AI alongside ML. Deadline typically in August; conference in February. Acceptance rate: ~20%. Strong venue for classical AI and applied/interdisciplinary work.
Tier 1.5: Highly Regarded Specialised Venues
IJCAI (International Joint Conference on Artificial Intelligence)
One of the oldest AI conferences (since 1969) and still prestigious, particularly in AI outside of deep learning — search, planning, knowledge-based systems, and robotics. Acceptance rate: ~13–15%, making it more selective than AAAI.
UAI (Conference on Uncertainty in Artificial Intelligence)
The premier venue for probabilistic and Bayesian AI methods. Acceptance rate ~25%. Strong community, tightly focused — if your paper is on probabilistic graphical models, causal inference, or Bayesian learning, UAI is your target.
AISTATS (International Conference on AI and Statistics)
The intersection of machine learning and statistics. Every paper needs both an algorithmic contribution and statistical grounding. Strong venue for researchers bridging these communities.
Specialised AI Subfield Conferences
- CVPR, ICCV, ECCV — Computer vision (see the CS conference guide)
- ACL, EMNLP, NAACL — Natural language processing
- ICRA, IROS — Robotics and autonomous systems
- KDD (Knowledge Discovery and Data Mining) — Applied data science and ML at scale
- CIKM (Information and Knowledge Management) — Information retrieval and applied AI
- ECML PKDD — European Machine Learning conference; excellent for European-based researchers
Emerging and Regional Venues Worth Watching
The AI conference landscape is expanding rapidly. Venues that have gained significant traction in recent years include:
- CoLLAs (Conference on Lifelong Learning Agents) — New but fast-growing community around continual learning
- COLM (Conference on Language Modeling) — Dedicated to LLM research; acceptance rate competitive
- TMLR (Transactions on Machine Learning Research) — A journal with rapid turnaround and open reviewing; increasingly cited alongside conferences
Should You Submit to Workshops?
Yes — strategically. Co-located workshops at NeurIPS, ICML, and ICLR attract hundreds of specialist attendees and are reviewed by field experts. A strong workshop paper, presented at the right venue, can generate more relevant discussion than a poster at a conference slightly outside your niche. Many papers start as workshop submissions and are refined into main-track papers at later editions.
Tips for First-Time AI Conference Submitters
- Post to arXiv the same day you submit — this is expected in the community and most conferences explicitly allow it
- Read 5–10 accepted papers from the previous edition of your target venue before writing; absorb the implicit style and depth expected
- Rebuttals matter at ICLR and NeurIPS — invest time in your author response, especially for borderline reviews
- The desk rejection rate at top venues for papers that are clearly out of scope can exceed 15%; submit to the right venue, not the most prestigious one that might accept you
Browse upcoming AI conference deadlines on LatestConferences.com and set up alerts so you never miss a submission window.