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 499 篇，第 4 / 10 页

  1. NeurIPS 2026Speed Predictions for Online Energy-Efficient Scheduling
  2. NeurIPS 2026Spike-SFT: Selective Parameter Enhancement and Fusion for Efficient Spiking Neural Networks
  3. NeurIPS 2026Step-dLLM: Adaptive Step-aware Sparse Attention for Efficient Diffusion LLM Inference
  4. NeurIPS 2026Stroke-Audited Quadratic Bezier Splatting with Structural Initialization for Efficient Painting Rendering
  5. NeurIPS 2026Structured Unitary Tensor Network Representations for Circuit-Efficient Quantum Data Encoding
  6. NeurIPS 2026Subdata Selection: A Unified Framework for Optimal Selection and Statistical Efficiency Assessment
  7. NeurIPS 2026SWE-Crafter: Scaling Executable Multilingual Software Engineering Data with Meta-Skill Agents
  8. NeurIPS 2026Systematic Scaling Analysis of Jailbreak Attacks in Large Language Models
  9. NeurIPS 2026Talk Less, Work More: Communication-Efficient Decentralized Stochastic Approximation
  10. NeurIPS 2026Target-Budget Best-Function Identification for Scaling-Law-Guided Model Selection
  11. NeurIPS 2026TeMPO: Frame-Causal Token Compression for Efficient Video Large Language Model
  12. NeurIPS 2026Temporal Smoothness Constraints on Efficient Neurobiological Codes Imply Temporal Specialization
  13. NeurIPS 2026The Dormant Spiking Neuron: A State-Driven Mechanism for Efficient Spiking Neural Networks
  14. NeurIPS 2026The Kernel Reality Check: Benchmarking and Distilling Efficient Attention at Scale
  15. NeurIPS 2026The SEC Filings Dataset: Reconstructing U.S. Corporate and Financial Disclosures into Layout-Faithful and Token-Efficient Pretraining Data
  16. NeurIPS 2026Token Filtering: Online Attention Pruning via KV Similarity for Efficient LLM Inference
  17. NeurIPS 2026TokenRouter: Efficient Serving System for Token-Level LLM Routing
  18. NeurIPS 2026ToLD: Efficient Time Series Forecasting via Tokenized Truncated Latent Diffusion
  19. NeurIPS 2026Towards Principled Efficient Rollout Allocation in Test-Time Training
  20. NeurIPS 2026TTB: Test-time MLP Baking for Efficient Rendering of Decoder-only View Synthesis Models
  21. NeurIPS 2026Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting
  22. NeurIPS 2026UTOPI: Efficient Egocentric Long-Video Understanding in AR via User-Guided Token Pre-Compression
  23. NeurIPS 2026Vendi Anomaly Scores for Efficient and Accurate Anomaly Detection
  24. NeurIPS 2026Vortex: Efficient and Programmable Sparse Attention Serving
  25. NeurIPS 2026What to Predict for Efficient Scheduling on Parallel Machines
  26. NeurIPS 2026Where You Backpropagate Matters: Token Hypothesis for Memory-Efficient Fine-Tuning
  27. NeurIPS 2026Winfree Oscillatory Neural Network
  28. NeurIPS 2025$\texttt{BetaConform}$: Efficient MAP Estimation of LLM Ensemble Judgment Performance with Prior Transfer
  29. NeurIPS 20254DGCPro: Efficient Hierarchical 4D Gaussian Compression for Progressive Volumetric Video Streaming
  30. NeurIPS 2025A Stable Whitening Optimizer for Efficient Neural Network Training
  31. NeurIPS 2025A Token is Worth over 1,000 Tokens: Efficient Knowledge Distillation through Low-Rank Clone
  32. NeurIPS 2025A Unified Framework for Provably Efficient Algorithms to Estimate Shapley Values
  33. NeurIPS 2025Accurate and Efficient Low-Rank Model Merging in Core Space
  34. NeurIPS 2025Accurate KV Cache Eviction via Anchor Direction Projection for Efficient LLM Inference
  35. NeurIPS 2025Active Measurement: Efficient Estimation at Scale
  36. NeurIPS 2025Active Seriation: Efficient Ordering Recovery with Statistical Guarantees
  37. NeurIPS 2025Ada-KV: Optimizing KV Cache Eviction by Adaptive Budget Allocation for Efficient LLM Inference
  38. NeurIPS 2025AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders
  39. NeurIPS 2025All You Need is One: Capsule Prompt Tuning with a Single Vector
  40. NeurIPS 2025AmorLIP: Efficient Language-Image Pretraining via Amortization
  41. NeurIPS 2025An Efficient Local Search Approach for Polarized Community Discovery in Signed Networks
  42. NeurIPS 2025Angles Don’t Lie: Unlocking Training‑Efficient RL Through the Model’s Own Signals
  43. NeurIPS 2025Are Language Models Efficient Reasoners? A Perspective from Logic Programming
  44. NeurIPS 2025Audits Under Resource, Data, and Access Constraints: Scaling Laws For Less Discriminatory Alternatives
  45. NeurIPS 2025Bilevel ZOFO: Efficient LLM Fine-Tuning and Meta-Training
  46. NeurIPS 2025Breaking the Compression Ceiling: Data-Free Pipeline for Ultra-Efficient Delta Compression
  47. NeurIPS 2025CAT: Circular-Convolutional Attention for Sub-Quadratic Transformers
  48. NeurIPS 2025Causal Spatio-Temporal Prediction: An Effective and Efficient Multi-Modal Approach
  49. NeurIPS 2025ChunkKV: Semantic-Preserving KV Cache Compression for Efficient Long-Context LLM Inference
  50. NeurIPS 2025COALA: Numerically Stable and Efficient Framework for Context-Aware Low-Rank Approximation

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