高效训练论文趋势 2023–2026

Scaling law、MoE、蒸馏与训练效率

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 1,765 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

高效训练论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
2023112678384346
2024207812280300
202519012611882516
2026177148169109603

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

全部高效训练论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 492 篇，第 4 / 10 页

  1. ICML 2026Swordsman: Entropy-Driven Adaptive Block Partition for Efficient Diffusion Language Models
  2. ICML 2026TabPack: Efficient Hyperparameter Ensembles for Tabular Deep Learning
  3. ICML 2026TetraJet-v2: Accurate NVFP4 Training for Large Language Models with Oscillation Suppression and Outlier Control
  4. ICML 2026ToaSt: Token Channel Selection and Structured Pruning for Efficient ViT
  5. ICML 2026Token Sparse Attention: Efficient Long-Context Inference with Interleaved Token Selection
  6. ICML 2026TokenDrop: Token-Level Importance-Aware Backward Propagation Skipping for Efficient LLM Fine-Tuning
  7. ICML 2026Towards Resource-Efficient LLMs: End-to-End Energy Accounting of Distillation Pipelines
  8. ICML 2026Training-Free Hashing-Based Attention via Binary Principal Components
  9. ICML 2026Turning Back Without Forgetting: Selective Backward Refinement for Parameter-Efficient Continual Learning
  10. ICML 2026Turning Stale Gradients into Stable Gradients: Coherent Coordinate Descent with Implicit Landscape Smoothing for Lightweight Zeroth-Order Optimization
  11. ICML 2026UniFast-HGR: Scalable and Efficient Maximal Correlation for Multimodal Models
  12. ICML 2026Unifying Dataset Pruning and Distillation for Efficient Large-scale Compression
  13. ICML 2026Unifying Stacking and Cascading for Efficient Ensemble Inference
  14. ICML 2026Universal One-third Time Scaling in Learning Peaked Distributions
  15. ICML 2026Untied Ulysses: Memory-Efficient Context Parallelism via Headwise Chunking
  16. ICML 2026Utility Boundary of Dataset Distillation: Scaling and Coverage Laws
  17. ICML 2026WAVE: Window-Aware Vocabulary-Efficient Early-Exit for Training-Free LLM Acceleration
  18. ICML 2026WBMM: Windowed Batch Matrix Multiplication for Efficient Large Receptive Field Convolution
  19. ICML 2026ZipMoE: Efficient On-Device MoE Serving via Lossless Compression and Cache-Affinity Scheduling
  20. ICML 2025A Hitchhiker's Guide to Scaling Law Estimation
  21. ICML 2025A Memory Efficient Randomized Subspace Optimization Method for Training Large Language Models
  22. ICML 2025A Sample Efficient Conditional Independence Test in the Presence of Discretization
  23. ICML 2025A Simple Model of Inference Scaling Laws
  24. ICML 2025Adapting Precomputed Features for Efficient Graph Condensation
  25. ICML 2025An Adaptive Orthogonal Convolution Scheme for Efficient and Flexible CNN Architectures
  26. ICML 2025An Efficient Matrix Multiplication Algorithm for Accelerating Inference in Binary and Ternary Neural Networks
  27. ICML 2025An Efficient Pruner for Large Language Model with Theoretical Guarantee
  28. ICML 2025An Efficient Search-and-Score Algorithm for Ancestral Graphs using Multivariate Information Scores for Complex Non-linear and Categorical Data
  29. ICML 2025An Expressive and Self-Adaptive Dynamical System for Efficient Function Learning
  30. ICML 2025Beyond Low-rank Decomposition: A Shortcut Approach for Efficient On-Device Learning
  31. ICML 2025Beyond Self-Repellent Kernels: History-Driven Target Towards Efficient Nonlinear MCMC on General Graphs
  32. ICML 2025BlockDialect: Block-wise Fine-grained Mixed Format Quantization for Energy-Efficient LLM Inference
  33. ICML 2025BSLoRA: Enhancing the Parameter Efficiency of LoRA with Intra-Layer and Inter-Layer Sharing
  34. ICML 2025COKE: Core Kernel for More Efficient Approximation of Kernel Weights in Multiple Kernel Clustering
  35. ICML 2025Complete-Tree Space Favors Data-Efficient Link Prediction
  36. ICML 2025Compositional Generalization via Forced Rendering of Disentangled Latents
  37. ICML 2025Connecting Thompson Sampling and UCB: Towards More Efficient Trade-offs Between Privacy and Regret
  38. ICML 2025Cost-efficient Collaboration between On-device and Cloud Language Models
  39. ICML 2025DA-KD: Difficulty-Aware Knowledge Distillation for Efficient Large Language Models
  40. ICML 2025Determinant Estimation under Memory Constraints and Neural Scaling Laws
  41. ICML 2025Distillation Scaling Laws
  42. ICML 2025Does Data Scaling Lead to Visual Compositional Generalization?
  43. ICML 2025EAGLES: Towards Effective, Efficient, and Economical Federated Graph Learning via Unified Sparsification
  44. ICML 2025Earley-Driven Dynamic Pruning for Efficient Structured Decoding
  45. ICML 2025Efficient and Separate Authentication Image Steganography Network
  46. ICML 2025Efficient ANN-SNN Conversion with Error Compensation Learning
  47. ICML 2025Efficient Bisection Projection to Ensure Neural-Network Solution Feasibility for Optimization over General Set
  48. ICML 2025Efficient Core-set Selection for Deep Learning Through Squared Loss Minimization
  49. ICML 2025Efficient Federated Incomplete Multi-View Clustering
  50. ICML 2025Efficient Graph Continual Learning via Lightweight Graph Neural Tangent Kernels-based Dataset Distillation

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 查高效训练在 NeurIPS 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": "efficient-training"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇高效训练论文？
共 1,765 篇，按年份分别是 2023 年 346 篇、2024 年 300 篇、2025 年 516 篇、2026 年 603 篇。
NeurIPS 2026 有多少篇高效训练论文？
NeurIPS 2026 有 177 篇高效训练论文，上一届是 190 篇。
ICLR 2026 有多少篇高效训练论文？
ICLR 2026 有 148 篇高效训练论文，上一届是 126 篇。
ICML 2026 有多少篇高效训练论文？
ICML 2026 有 169 篇高效训练论文，上一届是 118 篇。
CVPR 2026 有多少篇高效训练论文？
CVPR 2026 有 109 篇高效训练论文，上一届是 82 篇。
哪个会议的高效训练论文最多？
NeurIPS，2023–2026 年共 499 篇。

「效率与系统」下的其他主题