硬件协同论文趋势 2023–2026
加速器与编译
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 1,524 篇被归入该主题的论文。
主要结论
- 硬件协同论文从 2023 年的 173 篇增长到 2026 年的 765 篇，三年间扩大到约 4.4 倍。
- CVPR 是发硬件协同论文最多的会议，2023–2026 年共 479 篇，占四个会议总量的 31%。
- 交叉方向里涨得最快的是扩散模型：同时属于两个主题的论文从 2023 年的 4 篇增至 2026 年的 58 篇。
- 在「效率与系统」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的数据集与基准是 2.1 倍，硬件协同是 4.4 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 21 | 23 | 27 | 102 | 173 |
| 2024 | 69 | 45 | 55 | 94 | 263 |
| 2025 | 120 | 44 | 72 | 87 | 323 |
| 2026 | 263 | 120 | 186 | 196 | 765 |
代表论文
按年份和会议均匀挑选，方便快速看到每一届的样貌
- NeurIPS 2026AcceleGrad#: Adaptive Geometry-Aware Acceleration
- ICML 2026A Geometry-Aware Efficient Algorithm for Compositional Entropic Risk Minimization
- ICLR 2026Confident Block Diagonal Structure-Aware Invariable Graph Completion for Incomplete Multi-view Clustering
- CVPR 2026ActAvatar Temporally-Aware Precise Action Control for Talking Avatars
- NeurIPS 2025AcuRank: Uncertainty-Aware Adaptive Computation for Listwise Reranking
- ICML 2025CABS: Conflict-Aware and Balanced Sparsification for Enhancing Model Merging
- ICLR 2025FlashRNN: I/O-Aware Optimization of Traditional RNNs on modern hardware
- CVPR 2025A Distractor-Aware Memory for Visual Object Tracking with SAM2
- NeurIPS 2024HW-GPT-Bench: Hardware-Aware Architecture Benchmark for Language Models
- ICML 2024A Universal Class of Sharpness-Aware Minimization Algorithms
- ICLR 2024Domain-Inspired Sharpness-Aware Minimization Under Domain Shifts
- CVPR 2024Distribution-aware Knowledge Prototyping for Non-exemplar Lifelong Person Re-identification
- NeurIPS 2023Affinity-Aware Graph Networks
- ICML 2023A Three-regime Model of Network Pruning
- ICLR 2023$\rm A^2Q$: Aggregation-Aware Quantization for Graph Neural Networks
- CVPR 20233D-Aware Face Swapping
- NeurIPS 2026Accelerating Rectified Flow Models via Trajectory-Aware Caching
- ICML 2026Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse
- ICLR 2026Distribution-Aware Multi-Granularity Phase Coding: Towards Lower Conversion Error for Spike-Driven Large Language Models
- CVPR 2026BD-Merging Bias-Aware Dynamic Model Merging with Evidence-Guided Contrastive Learning
全部硬件协同论文
共 1,524 篇，第 31 / 31 页
- ICML 2023SpeedDETR: Speed-aware Transformers for End-to-end Object Detection
- ICML 2023Surrogate Module Learning: Reduce the Gradient Error Accumulation in Training Spiking Neural Networks
- ICML 2023Symmetry-Aware Robot Design with Structured Subgroups
- NeurIPS 2023A Robust and Opponent-Aware League Training Method for StarCraft II
- NeurIPS 2023Affinity-Aware Graph Networks
- NeurIPS 2023Coneheads: Hierarchy Aware Attention
- NeurIPS 2023Correlation Aware Sparsified Mean Estimation Using Random Projection
- NeurIPS 2023CSOT: Curriculum and Structure-Aware Optimal Transport for Learning with Noisy Labels
- NeurIPS 2023Environment-Aware Dynamic Graph Learning for Out-of-Distribution Generalization
- NeurIPS 2023Failure-Aware Gaussian Process Optimization with Regret Bounds
- NeurIPS 2023Fairness Aware Counterfactuals for Subgroups
- NeurIPS 2023Fine-Grained Cross-View Geo-Localization Using a Correlation-Aware Homography Estimator
- NeurIPS 2023HAP: Structure-Aware Masked Image Modeling for Human-Centric Perception
- NeurIPS 2023HotBEV: Hardware-oriented Transformer-based Multi-View 3D Detector for BEV Perception
- NeurIPS 2023Inner-Outer Aware Reconstruction Model for Monocular 3D Scene Reconstruction
- NeurIPS 2023Normalization Layers Are All That Sharpness-Aware Minimization Needs
- NeurIPS 2023Practical Sharpness-Aware Minimization Cannot Converge All the Way to Optima
- NeurIPS 2023Pruning vs Quantization: Which is Better?
- NeurIPS 2023Relative Entropic Optimal Transport: a (Prior-aware) Matching Perspective to (Unbalanced) Classification
- NeurIPS 2023Riemannian SAM: Sharpness-Aware Minimization on Riemannian Manifolds
- NeurIPS 2023SAME: Uncovering GNN Black Box with Structure-aware Shapley-based Multipiece Explanations
- NeurIPS 2023The Crucial Role of Normalization in Sharpness-Aware Minimization
- NeurIPS 2023The Grand Illusion: The Myth of Software Portability and Implications for ML Progress.
- NeurIPS 2023ZoomTrack: Target-aware Non-uniform Resizing for Efficient Visual Tracking
在 Claude、Cursor 里直接问这个趋势
接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。
可以直接复制这句话问 AI
用 paperlist 查硬件协同在 CVPR 2023–2026 的论文趋势，哪一年涨得最快？
AI 会调用的工具：get_topic_stats {"topic": "hardware-aware-ml"}
常见问题
- 2023–2026 年四大会议一共有多少篇硬件协同论文？
- 共 1,524 篇，按年份分别是 2023 年 173 篇、2024 年 263 篇、2025 年 323 篇、2026 年 765 篇。
- NeurIPS 2026 有多少篇硬件协同论文？
- NeurIPS 2026 有 263 篇硬件协同论文，上一届是 120 篇。
- ICLR 2026 有多少篇硬件协同论文？
- ICLR 2026 有 120 篇硬件协同论文，上一届是 44 篇。
- ICML 2026 有多少篇硬件协同论文？
- ICML 2026 有 186 篇硬件协同论文，上一届是 72 篇。
- CVPR 2026 有多少篇硬件协同论文？
- CVPR 2026 有 196 篇硬件协同论文，上一届是 87 篇。
- 哪个会议的硬件协同论文最多？
- CVPR，2023–2026 年共 479 篇。