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 603 篇，第 4 / 13 页

  1. ICLR 2026Efficient Ensemble Conditional Independence Test Framework for Causal Discovery
  2. ICLR 2026Efficient Learning on Large Graphs using a Densifying Regularity Lemma
  3. ICLR 2026Efficient Message-Passing Transformer for Error Correcting Codes
  4. ICLR 2026Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits
  5. ICLR 2026Efficient Orthogonal Fine-Tuning with Principal Subspace Adaptation
  6. ICLR 2026Efficient Prediction of Large Protein Complexes via Subunit-Guided Hierarchical Refinement
  7. ICLR 2026Efficient Quantization of Mixture-of-Experts with Theoretical Generalization Guarantees
  8. ICLR 2026Efficient Sliced Wasserstein Distance Computation via Adaptive Bayesian Optimization
  9. ICLR 2026Efficient-SAM2: Accelerating SAM2 with Object-Aware Visual Encoding and Memory Retrieval
  10. ICLR 2026EffiVMT: Video Motion Transfer via Efficient Spatial-Temporal Decoupled Finetuning
  11. ICLR 2026ELViS: Efficient Visual Similarity from Local Descriptors that Generalizes Across Domains
  12. ICLR 2026Energy-Efficient Random Variate Generation via Compressed Lookup Tables
  13. ICLR 2026Ensemble Prediction of Task Affinity for Efficient Multi-Task Learning
  14. ICLR 2026Ensembling Pruned Attention Heads For Uncertainty-Aware Efficient Transformers
  15. ICLR 2026Entropy-Based Block Pruning for Efficient Large Language Models
  16. ICLR 2026ES-dLLM: Efficient Inference for Diffusion Large Language Models by Early-Skipping
  17. ICLR 2026Expressive yet Efficient Feature Expansion with Adaptive Cross-Hadamard Products
  18. ICLR 2026Fast-dLLM v2: Efficient Block-Diffusion LLM
  19. ICLR 2026Fewer Battles, More Gain: An Information-Efficient Framework for Arena-based LLM Evaluation
  20. ICLR 2026Flatter Tokens are More Valuable for Speculative Draft Model Training
  21. ICLR 2026Fly-CL: A Fly-Inspired Framework for Enhancing Efficient Decorrelation and Reduced Training Time in Pre-trained Model-based Continual Representation Learning
  22. ICLR 2026Forge: Compiling a Unified Abstraction into Scalable Kernels for Linear Attention
  23. ICLR 2026FreeKV: Boosting KV Cache Retrieval for Efficient LLM Inference
  24. ICLR 2026FSA: An Alternative Efficient Implementation of Native Sparse Attention Kernel
  25. ICLR 2026FutureMind: Equipping Small Language Models with Strategic Thinking-Pattern Priors via Adaptive Knowledge Distillation
  26. ICLR 2026GAGA: Gaussianity-Aware Gaussian Approximation for Efficient 3D Molecular Generation
  27. ICLR 2026GradPruner: Gradient-guided Layer Pruning Enabling Efficient Fine-Tuning and Inference for LLMs
  28. ICLR 2026Grounding and Enhancing Informativeness and Utility in Dataset Distillation
  29. ICLR 2026Guided Speculative Inference for Efficient Test-Time Alignment of LLMs
  30. ICLR 2026HEAPr: Hessian-based Efficient Atomic Expert Pruning in Output Space
  31. ICLR 2026Highly Efficient and Effective LLMs with Multi-Boolean Architectures
  32. ICLR 2026Householder-Diagonalized Linear Attention (HDLA): Utilizing Enhanced Decay Mechanism for Efficient Sequence Modeling
  33. ICLR 2026How Reliable is Language Model Micro-Benchmarking?
  34. ICLR 2026How Text Quality Interventions Reshape Neural Scaling Laws for LLMs: Empirical Study
  35. ICLR 2026How to train data-efficient LLMs
  36. ICLR 2026Image is All You Need: Towards Efficient and Effective Large Language Model-Based Recommender Systems
  37. ICLR 2026Inconsistency Biases in Dynamic Data Pruning
  38. ICLR 2026KBVQ-MoE: KLT-guided SVD with Bias-Corrected Vector Quantization for MoE Large Language Models
  39. ICLR 2026Knowledge Distillation for Large Language Models through Residual Learning
  40. ICLR 2026Late-to-Early Training: LET LLMs Learn Earlier, So Faster and Better
  41. ICLR 2026Libra: Effective yet Efficient Load Balancing for Large-scale MoE Inference
  42. ICLR 2026LinearSR: Unlocking Linear Attention for Stable and Efficient Image Super-Resolution
  43. ICLR 2026LipNeXt: Scaling up Lipschitz-based Certified Robustness to Billion-parameter Models
  44. ICLR 2026LiteGuard: Efficient Task-Agnostic Model Fingerprinting with Enhanced Generalization
  45. ICLR 2026Lookup multivariate Kolmogorov-Arnold Networks
  46. ICLR 2026ManipEvalAgent: Promptable and Efficient Evaluation Framework for Robotic Manipulation Policies
  47. ICLR 2026MatRIS: Toward Reliable and Efficient Pretrained Machine Learning Interaction Potentials
  48. ICLR 2026MicroMix: Efficient Mixed-Precision Quantization with Microscaling Formats for Large Language Models
  49. ICLR 2026MiSS: Revisiting the Trade-off in LoRA with an Efficient Shard-Sharing Structure
  50. ICLR 2026MobiEdit: Resource-efficient Knowledge Editing for Personalized On-device LLMs

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