AAAI 2025 accepted papers by research area
3,486 papers were accepted at AAAI 2025 (+622 vs 2024). Here is how they split across 20 research areas, and which topics gained the most ground.
See AAAI 2025 in the interactive map
Research areas
| Area | Papers | Share | Change in share vs 2024 |
|---|---|---|---|
| Large language models | 501 | 14.4% | +6.0 pts |
| Label-efficient and robust learning | 394 | 11.3% | -2.6 pts |
| Vision-language and multimodal | 297 | 8.5% | +1.6 pts |
| Image generation | 291 | 8.3% | +0.7 pts |
| Learning theory and algorithms | 232 | 6.7% | -3.5 pts |
| Graph learning | 221 | 6.3% | -0.7 pts |
| Detection and segmentation | 181 | 5.2% | -0.1 pts |
| Video | 175 | 5.0% | +1.1 pts |
| Probabilistic and scientific modeling | 165 | 4.7% | -0.9 pts |
| Architectures and efficiency | 162 | 4.6% | -0.7 pts |
| Multi-agent systems and games | 140 | 4.0% | +0.8 pts |
| 3D vision | 140 | 4.0% | steady |
| Federated learning and privacy | 109 | 3.1% | steady |
| Reinforcement learning | 106 | 3.0% | -1.7 pts |
| Adversarial robustness and security | 103 | 3.0% | -0.3 pts |
| Point clouds | 75 | 2.2% | +0.3 pts |
| Time series | 73 | 2.1% | +0.5 pts |
| Continual learning | 49 | 1.4% | -0.3 pts |
| Optimization | 41 | 1.2% | steady |
| Bandits and online learning | 31 | 0.9% | -0.3 pts |
Fastest-growing topics at AAAI 2025
Change in share of AAAI papers from 2024 to 2025, topics with at least 10 papers.
- Language, safety and interpretability (Large language models): 86 → 154 papers
- Multimodal LLMs (Vision-language and multimodal): 7 → 55 papers
- Agents, code and math (Large language models): 32 → 77 papers
- Retrieval and knowledge (Large language models): 39 → 84 papers
- AI and society (Large language models): 61 → 108 papers
- Gaussian splatting (3D vision): 0 → 29 papers
- Image restoration (Image generation): 36 → 71 papers
- Instruction and fine-tuning (Large language models): 10 → 38 papers
- Alignment and preferences (Large language models): 2 → 24 papers
- Video generation (Video): 20 → 45 papers
More venues
Same year: ICML 2025 · NeurIPS 2025 · ICLR 2025 · CVPR 2025