上下文压缩论文趋势 2023–2026

主要结论

• 上下文压缩论文从 2023 年的 39 篇增长到 2026 年的 136 篇，三年间扩大到约 3.5 倍。
• NeurIPS 是发上下文压缩论文最多的会议，2023–2026 年共 123 篇，占四个会议总量的 34%。
• 交叉方向里涨得最快的是深度学习 · large language models：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 13 篇。
• 在「长上下文与记忆」的 5 个主题里，它的 2026 年论文数排第 4；同组增长最快的检索记忆是 6.6 倍，上下文压缩是 3.5 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 Free Lunch in LLM Compression: Revisiting Retraining after Pruning
• ICML 2026A Unified Sparse Attention via Multi-Granularity Compression
• ICLR 2026An evolutionary perspective on modes of learning in Transformers
• CVPR 2026AstraNav-Memory Contexts Compression for Long Memory
• NeurIPS 2025AdmTree: Compressing Lengthy Context with Adaptive Semantic Trees
• ICML 2025Does learning the right latent variables necessarily improve in-context learning?
• ICLR 2025Approaching Rate-Distortion Limits in Neural Compression with Lattice Transform Coding
• CVPR 2025ECVC: Exploiting Non-Local Correlations in Multiple Frames for Contextual Video Compression
• NeurIPS 2024StreamingDialogue: Prolonged Dialogue Learning via Long Context Compression with Minimal Losses
• ICML 2024Agnostic Sample Compression Schemes for Regression
• ICLR 2024Compressed Context Memory for Online Language Model Interaction
• CVPR 2024Diversity-aware Channel Pruning for StyleGAN Compression
• NeurIPS 2023Do Not Marginalize Mechanisms, Rather Consolidate!
• ICML 2023Private Federated Learning with Autotuned Compression
• ICLR 2023$k$NN Prompting: Beyond-Context Learning with Calibration-Free Nearest Neighbor Inference
• CVPR 2023AutoAD: Movie Description in Context
• NeurIPS 2026A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression
• ICML 2026Context-Driven Incremental Compression for Multi-Turn Dialogue Generation
• ICLR 2026Autoencoding-Free Context Compression for LLMs via Contextual Semantic Anchors
• CVPR 2026Featurising Pixels from Dynamic 3D Scenes with Linear In-Context Learners

全部上下文压缩论文

共 81 篇，第 1 / 2 页

1. CVPR 2024Amodal Completion via Progressive Mixed Context Diffusion
2. CVPR 2024CLiC: Concept Learning in Context
3. CVPR 2024ContextSeg: Sketch Semantic Segmentation by Querying the Context with Attention
4. CVPR 2024Diversity-aware Channel Pruning for StyleGAN Compression
5. CVPR 2024In-Context Matting
6. CVPR 2024Learned Lossless Image Compression based on Bit Plane Slicing
7. ICLR 2024ACRF: Compressing Explicit Neural Radiance Fields via Attribute Compression
8. ICLR 2024CausalLM is not optimal for in-context learning
9. ICLR 2024Compressed Context Memory for Online Language Model Interaction
10. ICLR 2024Context is Environment
11. ICLR 2024Context-Aware Meta-Learning
12. ICLR 2024How do Language Models Bind Entities in Context?
13. ICLR 2024HyperAttention: Long-context Attention in Near-Linear Time
14. ICLR 2024IDEAL: Influence-Driven Selective Annotations Empower In-Context Learners in Large Language Models
15. ICLR 2024In defense of parameter sharing for model-compression
16. ICLR 2024In-context Autoencoder for Context Compression in a Large Language Model
17. ICLR 2024In-Context Learning Learns Label Relationships but Is Not Conventional Learning
18. ICLR 2024Information Bottleneck Analysis of Deep Neural Networks via Lossy Compression
19. ICLR 2024Language Modeling Is Compression
20. ICLR 2024RECOMP: Improving Retrieval-Augmented LMs with Context Compression and Selective Augmentation
21. ICLR 2024Towards image compression with perfect realism at ultra-low bitrates
22. ICML 2024Agnostic Sample Compression Schemes for Regression
23. ICML 2024An Information-Theoretic Analysis of In-Context Learning
24. ICML 2024Breaking through the learning plateaus of in-context learning in Transformer
25. ICML 2024Can Mamba Learn How To Learn? A Comparative Study on In-Context Learning Tasks
26. ICML 2024Conditional Language Learning with Context
27. ICML 2024Data-free Neural Representation Compression with Riemannian Neural Dynamics
28. ICML 2024Debiased Distribution Compression
29. ICML 2024Dual Operating Modes of In-Context Learning
30. ICML 2024Exact Conversion of In-Context Learning to Model Weights in Linearized-Attention Transformers
31. ICML 2024ExCP: Extreme LLM Checkpoint Compression via Weight-Momentum Joint Shrinking
32. ICML 2024Extreme Compression of Large Language Models via Additive Quantization
33. ICML 2024Feedback Loops With Language Models Drive In-Context Reward Hacking
34. ICML 2024Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference
35. ICML 2024How do Transformers Perform In-Context Autoregressive Learning ?
36. ICML 2024In-context Convergence of Transformers
37. ICML 2024In-Context Learning Agents Are Asymmetric Belief Updaters
38. ICML 2024In-context Learning on Function Classes Unveiled for Transformers
39. ICML 2024In-Context Principle Learning from Mistakes
40. ICML 2024In-Context Unlearning: Language Models as Few-Shot Unlearners
41. ICML 2024Is In-Context Learning in Large Language Models Bayesian? A Martingale Perspective
42. ICML 2024LLM Maybe LongLM: SelfExtend LLM Context Window Without Tuning
43. ICML 2024LoCoCo: Dropping In Convolutions for Long Context Compression
44. ICML 2024LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens
45. ICML 2024LoRAP: Transformer Sub-Layers Deserve Differentiated Structured Compression for Large Language Models
46. ICML 2024Neural NeRF Compression
47. ICML 2024PICLe: Eliciting Diverse Behaviors from Large Language Models with Persona In-Context Learning
48. ICML 2024Position: Do pretrained Transformers Learn In-Context by Gradient Descent?
49. ICML 2024Rethinking Adversarial Robustness in the Context of the Right to be Forgotten
50. ICML 2024Training-Free Long-Context Scaling of Large Language Models

常见问题

2023–2026 年四大会议一共有多少篇上下文压缩论文？

共 365 篇，按年份分别是 2023 年 39 篇、2024 年 81 篇、2025 年 109 篇、2026 年 136 篇。

NeurIPS 2026 有多少篇上下文压缩论文？

NeurIPS 2026 有 50 篇上下文压缩论文，上一届是 38 篇。

ICLR 2026 有多少篇上下文压缩论文？

ICLR 2026 有 28 篇上下文压缩论文，上一届是 32 篇。

ICML 2026 有多少篇上下文压缩论文？

ICML 2026 有 39 篇上下文压缩论文，上一届是 20 篇。

CVPR 2026 有多少篇上下文压缩论文？

CVPR 2026 有 19 篇上下文压缩论文，上一届是 19 篇。

哪个会议的上下文压缩论文最多？

NeurIPS，2023–2026 年共 123 篇。