缩放定律论文趋势 2023–2026

主要结论

• 缩放定律论文从 2023 年的 60 篇增长到 2026 年的 248 篇，三年间扩大到约 4.1 倍。
• NeurIPS 是发缩放定律论文最多的会议，2023–2026 年共 221 篇，占四个会议总量的 37%。
• 交叉方向里涨得最快的是时间序列：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 24 篇。
• 在「优化与缩放」的 5 个主题里，它的 2026 年论文数排第 5；同组增长最快的低秩适配是 3.0 倍，缩放定律是 4.1 倍。

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 2026A Unified Merge Calculus for Learning-Rate Scaling on Neural Computation Graphs
• ICML 2026A Capacity-Based Rationale for Multi-Head Attention
• ICLR 2026$\mathbf{Li_2}$: A Framework on Dynamics of Feature Emergence and Delayed Generalization
• CVPR 2026A Unified Framework for Knowledge Transfer in Bidirectional Model Scaling
• NeurIPS 2025Audits Under Resource, Data, and Access Constraints: Scaling Laws For Less Discriminatory Alternatives
• ICML 2025A Simple Model of Inference Scaling Laws
• ICLR 2025(Mis)Fitting Scaling Laws: A Survey of Scaling Law Fitting Techniques in Deep Learning
• CVPR 2025D^3: Scaling Up Deepfake Detection by Learning from Discrepancy
• NeurIPS 2024An exactly solvable model for emergence and scaling laws in the multitask sparse parity problem
• ICML 2024A Dynamical Model of Neural Scaling Laws
• ICLR 2024Principled Architecture-aware Scaling of Hyperparameters
• CVPR 2024From Activation to Initialization: Scaling Insights for Optimizing Neural Fields
• NeurIPS 2023Abide by the law and follow the flow: conservation laws for gradient flows
• ICML 2023Dissecting the Effects of SGD Noise in Distinct Regimes of Deep Learning
• ICLR 2023Distributionally Robust Post-hoc Classifiers under Prior Shifts
• CVPR 2023LargeKernel3D: Scaling Up Kernels in 3D Sparse CNNs
• NeurIPS 2026ADIS-Law: Unified Scaling Laws for Annealing-Phase Domain Injection in Large Language Models
• ICML 2026Balancing Learning Rates Across Layers: Exact Two-Step Dynamics and Optimal Scaling in Linear Neural Networks
• ICLR 2026A universal compression theory: Lottery ticket hypothesis and superpolynomial scaling laws
• CVPR 2026MatAnyone 2 Scaling Video Matting via a Learned Quality Evaluator

全部缩放定律论文

共 78 篇，第 1 / 2 页

1. ICML 2026$L^3$: Large Lookup Layers
2. ICML 2026A Capacity-Based Rationale for Multi-Head Attention
3. ICML 2026Active Budget Allocation for Efficient Scaling Law Estimation via Surrogate-Guided Pruning
4. ICML 2026Balancing Learning Rates Across Layers: Exact Two-Step Dynamics and Optimal Scaling in Linear Neural Networks
5. ICML 2026Branch Scaling Manifests as Implicit Architectural Regularization for Improving Generalization in Overparameterized ResNets
6. ICML 2026cMoLLM at Scale: Horizontal Scaling Laws for Convolutionally-Gated Mixture-of-LLMs
7. ICML 2026Conservation Laws for Modern Neural Architectures
8. ICML 2026Curriculum-Guided Layer Scaling for Language Model Pretraining
9. ICML 2026Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training
10. ICML 2026Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling
11. ICML 2026Deriving Neural Scaling Laws from the Statistics of Natural Language
12. ICML 2026Discovering Scaling Exponents with Physics-Informed Müntz-Szász Networks
13. ICML 2026DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Use
14. ICML 2026dnaHNet: A Scalable and Hierarchical Foundation Model for Genomic Sequence Learning
15. ICML 2026DOCKSMITH: Scaling Reliable Coding Environments via an Agentic Docker Builder
16. ICML 2026Dropout Universality: Scaling Laws and Optimal Scheduling at the Edge-of-Chaos
17. ICML 2026Dynamics of neural scaling laws in random feature regression with powerlaw-distributed kernel eigenvalues
18. ICML 2026ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment
19. ICML 2026Expected Return Causes Outcome-Level Mode Collapse in Reinforcement Learning and How to Fix It with Inverse Probability Scaling
20. ICML 2026Explaining Data Mixing Scaling Laws
21. ICML 2026Finding the Minimal Parameter Budget for Implicit Reasoning: A Data Complexity Driven Scaling Law for Language Models
22. ICML 2026FUSE: Ensembling Verifiers with Zero Labeled Data
23. ICML 2026Generalization and Scaling Laws for Mixture-of-ExpertsTransformers
24. ICML 2026How does Chain of Thought decompose complex tasks?
25. ICML 2026How Many Different Outputs Can a Transformer Generate?
26. ICML 2026Hyperparameter Transfer Laws for Non-Recurrent Multi-Path Neural Networks
27. ICML 2026Improved Scaling Laws via Weak-to-Strong Generalization in Random Features Ridge Regression
28. ICML 2026InfoLaw: Information Scaling Laws for Large Language Models with Quality-Weighted Mixture Data and Repetition
29. ICML 2026Inverse Depth Scaling From Most Layers Being Similar
30. ICML 2026IsoCompute Playbook: Optimally Scaling Sampling Compute for LLM RL
31. ICML 2026Item Response Scaling Laws: A Measurement Theory Approach for Efficient and Generalizable Neural Scaling Estimation
32. ICML 2026LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling Laws
33. ICML 2026Making Models Unmergeable via Scaling-Sensitive Loss Landscape
34. ICML 2026Model Merging Scaling Laws in Large Language Models
35. ICML 2026MOOSE-Star: Unlocking Tractable Training for Scientific Discovery by Breaking the Complexity Barrier
36. ICML 2026Muon in Associative Memory Learning: Training Dynamics and Scaling Laws
37. ICML 2026Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings
38. ICML 2026On the Origin of Neural Scaling Laws: from Random Graphs to Natural Language
39. ICML 2026On the Power of (Approximate) Reward Models for Inference-Time Scaling: Sequential Monte Carlo and Beyond
40. ICML 2026Optimal Splitting of Language Models from Mixtures to Specialized Domains
41. ICML 2026OXE-AugE: A Large-Scale Robot Augmentation of OXE for Scaling Cross-Embodiment Policy Learning
42. ICML 2026Particles Don’t Care About Z: Towards Scaling Entropy Estimation of Unnormalized Densities
43. ICML 2026PluRel: Synthetic Data unlocks Scaling Laws for Relational Foundation Models
44. ICML 2026POET-X: Memory-efficient LLM Training by Scaling Orthogonal Transformation
45. ICML 2026Predicting Large Model Test Losses with a Noisy Quadratic System
46. ICML 2026Prescriptive Scaling Reveals the Evolution of Language Model Capabilities
47. ICML 2026Principled Synthetic Data Enables the First Scaling Laws for LLMs in Recommendation
48. ICML 2026Random Scaling of Emergent Capabilities
49. ICML 2026Resolving the Timestep Scaling Paradox in Spiking Neural Networks with a Timestep-Scalable Neuron Model
50. ICML 2026Rethinking Generative Image Pretraining: How Far Are We From Scaling Up Next-Pixel Prediction?

常见问题

2023–2026 年四大会议一共有多少篇缩放定律论文？

共 605 篇，按年份分别是 2023 年 60 篇、2024 年 96 篇、2025 年 201 篇、2026 年 248 篇。

NeurIPS 2026 有多少篇缩放定律论文？

NeurIPS 2026 有 82 篇缩放定律论文，上一届是 80 篇。

ICLR 2026 有多少篇缩放定律论文？

ICLR 2026 有 73 篇缩放定律论文，上一届是 48 篇。

ICML 2026 有多少篇缩放定律论文？

ICML 2026 有 78 篇缩放定律论文，上一届是 56 篇。

CVPR 2026 有多少篇缩放定律论文？

CVPR 2026 有 15 篇缩放定律论文，上一届是 17 篇。

哪个会议的缩放定律论文最多？

NeurIPS，2023–2026 年共 221 篇。