NeurIPS 2017 accepted papers by research area
688 papers were accepted at NeurIPS 2017 (+118 vs 2016). Here is how they split across 20 research areas, and which topics gained the most ground.
See NeurIPS 2017 in the interactive map
Submitting to NeurIPS? Checklist →
Research areas
| Area | Papers | Share | Change in share vs 2016 |
|---|---|---|---|
| Learning theory and algorithms | 188 | 27.3% | -2.8 pts |
| Probabilistic and scientific modeling | 127 | 18.5% | -0.3 pts |
| Optimization | 71 | 10.3% | -0.7 pts |
| Architectures and efficiency | 57 | 8.3% | +0.9 pts |
| Label-efficient and robust learning | 46 | 6.7% | +0.4 pts |
| Bandits and online learning | 42 | 6.1% | -0.4 pts |
| Reinforcement learning | 36 | 5.2% | +1.4 pts |
| Image generation | 29 | 4.2% | +2.1 pts |
| Graph learning | 19 | 2.8% | +0.1 pts |
| Multi-agent systems and games | 13 | 1.9% | +0.8 pts |
| Vision-language and multimodal | 11 | 1.6% | -0.2 pts |
| Detection and segmentation | 8 | 1.2% | -0.8 pts |
| Video | 8 | 1.2% | steady |
| Time series | 8 | 1.2% | -0.2 pts |
| Federated learning and privacy | 7 | 1.0% | +0.3 pts |
| 3D vision | 5 | 0.7% | -1.0 pts |
| Adversarial robustness and security | 5 | 0.7% | steady |
| Continual learning | 4 | 0.6% | +0.6 pts |
| Large language models | 2 | 0.3% | -0.2 pts |
| Point clouds | 2 | 0.3% | +0.1 pts |
Fastest-growing topics at NeurIPS 2017
Change in share of NeurIPS papers from 2016 to 2017, topics with at least 10 papers.
- Kernels and regression theory (Learning theory and algorithms): 11 → 23 papers
- GANs (Image generation): 3 → 13 papers
- Policy gradient and actor-critic (Reinforcement learning): 6 → 16 papers
- Uncertainty and Gaussian processes (Probabilistic and scientific modeling): 16 → 24 papers
- Dynamical systems (Probabilistic and scientific modeling): 15 → 22 papers
- RNNs and normalization (Architectures and efficiency): 23 → 31 papers
- Graph algorithms and clustering (Graph learning): 11 → 13 papers
- Online convex optimization and games (Bandits and online learning): 18 → 21 papers
- Visual and representation RL (Reinforcement learning): 10 → 11 papers
- Molecules and proteins (Probabilistic and scientific modeling): 14 → 15 papers
More venues
Same year: ICML 2017 · ICLR 2017 · CVPR 2017 · AAAI 2017 · ICCV 2017 · TPAMI 2017 · IJCV 2017 · JMLR 2017
NeurIPS in other years: 2016 · 2018 · 2019 · 2020 · 2021 · 2022 · 2023 · 2024 · 2025 · 2026
