界面理解论文趋势 2023–2026

主要结论

• 界面理解论文从 2023 年的 15 篇增长到 2026 年的 27 篇，三年间扩大到约 1.8 倍。
• NeurIPS 是发界面理解论文最多的会议，2023–2026 年共 23 篇，占四个会议总量的 30%。
• 交叉方向里涨得最快的是AI 智能体：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 9 篇。
• 在「教育与交互」的 5 个主题里，它的 2026 年论文数排第 3；同组增长最快的用户建模是 2.9 倍，界面理解是 1.8 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 2026Denoise First, Orthogonalize Later: Understanding Momentum in Muon via Spectral Filtering
• ICML 2026A Single Layer to Explain Them All: Understanding Massive Values in Large Language Models
• ICLR 2026Towards Understanding the Shape of Representations in Protein Language Models
• CVPR 2026RISE Single Static Radar-based Indoor Scene Understanding
• NeurIPS 2025The Graphon Limit Hypothesis: Understanding Neural Network Pruning via Infinite Width Analysis
• ICML 2025On Understanding Attention-Based In-Context Learning for Categorical Data
• ICLR 2025Toward Understanding In-context vs. In-weight Learning
• CVPR 2025MP-GUI: Modality Perception with MLLMs for GUI Understanding
• ICML 2024CompeteAI: Understanding the Competition Dynamics of Large Language Model-based Agents
• ICLR 2024Understanding In-Context Learning from Repetitions
• CVPR 2024PanoContext-Former: Panoramic Total Scene Understanding with a Transformer
• NeurIPS 2023Understanding the detrimental class-level effects of data augmentation
• ICML 2023How Do Transformers Learn Topic Structure: Towards a Mechanistic Understanding
• ICLR 2023First Steps Toward Understanding the Extrapolation of Nonlinear Models to Unseen Domains
• CVPR 2023AssemblyHands: Towards Egocentric Activity Understanding via 3D Hand Pose Estimation
• NeurIPS 2026Shifting the Gradient: Understanding How Defensive Training Methods Protect Language Model Integrity
• ICML 2026Understanding the Parameter Space Geometry of Transformers Encoding Boolean Functions
• ICLR 2026Understanding In-Context Learning on Structured Manifolds: Bridging Attention to Kernel Methods
• CVPR 20264D-RGPT Toward Region-level 4D Understanding via Perceptual Distillation
• NeurIPS 2025Towards a Geometric Understanding of Tensor Learning via the t-Product

全部界面理解论文

共 20 篇，第 1 / 1 页

1. ICML 2026A Single Layer to Explain Them All: Understanding Massive Values in Large Language Models
2. ICML 2026Understanding the Parameter Space Geometry of Transformers Encoding Boolean Functions
3. ICML 2025On Understanding Attention-Based In-Context Learning for Categorical Data
4. ICML 2025Position: AI Agents Need Authenticated Delegation
5. ICML 2025Understanding Mode Connectivity via Parameter Space Symmetry
6. ICML 2025Understanding Nonlinear Implicit Bias via Region Counts in Input Space
7. ICML 2025Understanding the Kronecker Matrix-Vector Complexity of Linear Algebra
8. ICML 2024BAGEL: Bootstrapping Agents by Guiding Exploration with Language
9. ICML 2024CompeteAI: Understanding the Competition Dynamics of Large Language Model-based Agents
10. ICML 2024DoraemonGPT: Toward Understanding Dynamic Scenes with Large Language Models (Exemplified as A Video Agent)
11. ICML 2024Dr. Strategy: Model-Based Generalist Agents with Strategic Dreaming
12. ICML 2024InfiAgent-DABench: Evaluating Agents on Data Analysis Tasks
13. ICML 2024LangCell: Language-Cell Pre-training for Cell Identity Understanding
14. ICML 2024MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation
15. ICML 2024Towards a Better Theoretical Understanding of Independent Subnetwork Training
16. ICML 2024Towards Theoretical Understanding of Learning Large-scale Dependent Data via Random Features
17. ICML 2024Towards Understanding Inductive Bias in Transformers: A View From Infinity
18. ICML 2024Understanding the Effects of Iterative Prompting on Truthfulness
19. ICML 2023How Do Transformers Learn Topic Structure: Towards a Mechanistic Understanding
20. ICML 2023Understanding Incremental Learning of Gradient Descent: A Fine-grained Analysis of Matrix Sensing

常见问题

2023–2026 年四大会议一共有多少篇界面理解论文？

共 77 篇，按年份分别是 2023 年 15 篇、2024 年 16 篇、2025 年 19 篇、2026 年 27 篇。

NeurIPS 2026 有多少篇界面理解论文？

NeurIPS 2026 有 16 篇界面理解论文，上一届是 6 篇。

ICLR 2026 有多少篇界面理解论文？

ICLR 2026 有 2 篇界面理解论文，上一届是 2 篇。

ICML 2026 有多少篇界面理解论文？

ICML 2026 有 2 篇界面理解论文，上一届是 5 篇。

CVPR 2026 有多少篇界面理解论文？

CVPR 2026 有 7 篇界面理解论文，上一届是 6 篇。

哪个会议的界面理解论文最多？

NeurIPS，2023–2026 年共 23 篇。