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 10 / 13 页

  1. NeurIPS 2026Communication-Efficient Personalized Adaptation via Federated-Local Model Merging
  2. NeurIPS 2026CommunityKV: Efficient Long-Context Decoding via Graph Partitioning
  3. NeurIPS 2026CompactAttention: Accelerating Chunked Prefill with Block-Union KV Selection
  4. NeurIPS 2026Confidence-Based Decoding is Provably Efficient for Diffusion Language Models
  5. NeurIPS 2026Constrained Decoding for Diffusion Language Models via Efficient Inference over Finite Automata
  6. NeurIPS 2026CoPE-VideoLM: Leveraging Codec Primitives For Efficient Video Language Modelling
  7. NeurIPS 2026CRUMB: Efficient Prior Fitted Network Inference via Distributionally Matched Context Batching
  8. NeurIPS 2026Data-Constrained Language Model Pretraining: Improved Regularization and Scaling Laws
  9. NeurIPS 2026Data-Free Reservoir Features for Efficient Long-Horizon Cold-Start Continual Learning
  10. NeurIPS 2026Dataset Distillation via Drifting
  11. NeurIPS 2026Decomposing and Reshaping Scaling Laws through Token Learning Times
  12. NeurIPS 2026Depth through Recurrence: Towards Ultra-Efficient On-Device ASR
  13. NeurIPS 2026DrugSAGE: Self-evolving Agent Experience for Efficient State-of-the-Art Drug Discovery
  14. NeurIPS 2026DualKV: Shared-Prompt Flash Attention for Efficient RL Training with Large Rollouts and Long Contexts
  15. NeurIPS 2026Dynamic Convolutions Improve Transformers
  16. NeurIPS 2026Dynamic Resolution Routing for Efficient Egocentric Grounding
  17. NeurIPS 2026Dynamics-Aware Sparse Attention for Efficient Autoregressive Video Diffusion
  18. NeurIPS 2026DyRA: Dynamic Residual Approximation for Efficient Matrix Multiplication in DNNs
  19. NeurIPS 2026Efficient Algorithms for Distributed Saddle Problems
  20. NeurIPS 2026Efficient Algorithms For Fully Dynamic Bipartite Matching In Metric Spaces
  21. NeurIPS 2026Efficient algorithms for linear regression with heteroskedastic errors
  22. NeurIPS 2026Efficient Benchmarking Is Just Feature Selection and Multiple Regression
  23. NeurIPS 2026Efficient Brain-to-Speech Decoding with Fixed-Delay Spiking Neural Networks
  24. NeurIPS 2026Efficient Collaborative LLM Fine-Tuning over Heterogeneous Mobile Devices via Many Backbones to One Side-Network Tuning
  25. NeurIPS 2026Efficient Computation and Best-Response Dynamics in Anonymous Two-Action Games with Linear Utilities
  26. NeurIPS 2026Efficient Decoder Scaling Strategy for Constructive Neural Routing Solvers
  27. NeurIPS 2026Efficient Fine-Tuning for Structured Sparsity Under Group Repartitioning
  28. NeurIPS 2026Efficient Image Synthesis with Sphere Latent Encoder
  29. NeurIPS 2026Efficient Label Distribution Inference Attack and Defense on Classifier Weights in Federated Learning
  30. NeurIPS 2026Efficient Learning of Truncated Boolean Product Distributions: Influence to the Rescue
  31. NeurIPS 2026Efficient LLM Adaptation with Forward-Only Passes
  32. NeurIPS 2026Efficient Memory Crystallization for Graph Learning under Non-Stationary Distribution Shifts
  33. NeurIPS 2026Efficient Pre-Training with Token Superposition
  34. NeurIPS 2026Efficient Scaling of LLM Training with Flexible Context Parallelism
  35. NeurIPS 2026Efficient Serving for Dynamic Agent Workflows with Prediction-based KV-Cache Management
  36. NeurIPS 2026Efficient Training of Deep Spiking Neural Networks with Input-Driven Derivative-Free Updates
  37. NeurIPS 2026Efficient Variational Inference for Log-Gaussian Cox Processes via Voronoi Tessellation
  38. NeurIPS 2026Efficient, Accurate and Stable Gradients for Neural ODEs
  39. NeurIPS 2026Efficiently Aligning Draft Models via Parameter- and Data-Efficient Adaptation
  40. NeurIPS 2026EngramState: Loadable Tool Priors for Efficient Function Calling
  41. NeurIPS 2026EventPrune: Cascaded Event-Assisted Token Pruning for Efficient First-Person Dynamic Spatial Reasoning
  42. NeurIPS 2026Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction
  43. NeurIPS 2026FELPS: Fair and Efficient Scheduling for Multi-LoRA Serving System
  44. NeurIPS 2026Finite-Memory Control of POMDPs: Fundamental Limits and Efficient Design
  45. NeurIPS 2026Finite-Sample and Communication-Efficient Networked Information Aggregation
  46. NeurIPS 2026Flash-KMeans: Fast and Memory-Efficient Exact K-Means
  47. NeurIPS 2026FlashMask-3: Efficient and Expressive Mask-Aware Distributed Attention
  48. NeurIPS 2026Flow-Direct: Feedback-Efficient and Reusable Guidance for Flow Models via Non-Parametric Guidance Field
  49. NeurIPS 2026From Experts to Sub-experts: Fine-grained Parameter-Efficient Fine-Tuning for MoE LLMs
  50. NeurIPS 2026Generative Cross-Entropy: A Strictly Proper Loss for Data-Efficient Classification

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