视听生成论文趋势 2023–2026

主要结论

• 视听生成论文从 2023 年的 92 篇增长到 2026 年的 336 篇，三年间扩大到约 3.7 倍。
• CVPR 是发视听生成论文最多的会议，2023–2026 年共 383 篇，占四个会议总量的 49%。
• 交叉方向里涨得最快的是多模态大模型：同时属于两个主题的论文从 2023 年的 6 篇增至 2026 年的 79 篇。
• 在「多模态生成」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的叙事生成是 15 倍，视听生成是 3.7 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 2026Binding Visual Features Point by Point
• ICML 2026AVI-Bench: Toward Human-like Audio-Visual Intelligence of Omni-MLLMs
• ICLR 2026AC-Foley: Reference-Audio-Guided Video-to-Audio Synthesis with Acoustic Transfer
• CVPR 2026Benchmarking Single-Factor Physical Video-to-Audio Generation
• NeurIPS 2025$\texttt{AVROBUSTBENCH}$: Benchmarking the Robustness of Audio-Visual Recognition Models at Test-Time
• ICML 2025Fundamental Limits of Visual Autoregressive Transformers: Universal Approximation Abilities
• ICLR 2025AVHBench: A Cross-Modal Hallucination Benchmark for Audio-Visual Large Language Models
• CVPR 2025Adaptive Markup Language Generation for Contextually-Grounded Visual Document Understanding
• NeurIPS 2024Benchmarking Generative Models on Computational Thinking Tests in Elementary Visual Programming
• ICML 2024CaRiNG: Learning Temporal Causal Representation under Non-Invertible Generation Process
• ICLR 2024Can Transformers Capture Spatial Relations between Objects?
• CVPR 2024EmoVIT: Revolutionizing Emotion Insights with Visual Instruction Tuning
• NeurIPS 20234D Panoptic Scene Graph Generation
• ICML 2023A Unified Audio-Visual Learning Framework for Localization, Separation, and Recognition
• ICLR 2023Learning to reason over visual objects
• CVPR 2023Autoregressive Visual Tracking
• NeurIPS 2026From Articulated Kinematics to Routed Visual Control for Action-Conditioned Surgical Video Generation
• ICML 2026Deep Ensemble Clustering for Visual Representation Learning
• ICLR 2026AQuA: Toward Strategic Response Generation for Ambiguous Visual Questions
• CVPR 2026Co-Me Confidence Guided Token Merging for Visual Geometric Transformers

全部视听生成论文

共 19 篇，第 1 / 1 页

1. ICML 2025Continuous Visual Autoregressive Generation via Score Maximization
2. ICML 2025Contrastive Visual Data Augmentation
3. ICML 2025DCBM: Data-Efficient Visual Concept Bottleneck Models
4. ICML 2025Fundamental Limits of Visual Autoregressive Transformers: Universal Approximation Abilities
5. ICML 2025Learnings from Scaling Visual Tokenizers for Reconstruction and Generation
6. ICML 2025OmniAudio: Generating Spatial Audio from 360-Degree Video
7. ICML 2025Probing Visual Language Priors in VLMs
8. ICML 2025SketchDNN: Joint Continuous-Discrete Diffusion for CAD Sketch Generation
9. ICML 2025Sounding that Object: Interactive Object-Aware Image to Audio Generation
10. ICML 2025TabNAT: A Continuous-Discrete Joint Generative Framework for Tabular Data
11. ICML 2025Text-to-CAD Generation Through Infusing Visual Feedback in Large Language Models
12. ICML 2025The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models Via Visual Information Steering
13. ICML 2025Token Coordinated Prompt Attention is Needed for Visual Prompting
14. ICML 2025Understanding Complexity in VideoQA via Visual Program Generation
15. ICML 2025Vector Grimoire: Codebook-based Shape Generation under Raster Image Supervision
16. ICML 2025video-SALMONN-o1: Reasoning-enhanced Audio-visual Large Language Model
17. ICML 2025Visual Attention Never Fades: Selective Progressive Attention ReCalibration for Detailed Image Captioning in Multimodal Large Language Models
18. ICML 2025Visual Generation Without Guidance
19. ICML 2025ZipAR: Parallel Autoregressive Image Generation through Spatial Locality

常见问题

2023–2026 年四大会议一共有多少篇视听生成论文？

共 780 篇，按年份分别是 2023 年 92 篇、2024 年 131 篇、2025 年 221 篇、2026 年 336 篇。

NeurIPS 2026 有多少篇视听生成论文？

NeurIPS 2026 有 84 篇视听生成论文，上一届是 57 篇。

ICLR 2026 有多少篇视听生成论文？

ICLR 2026 有 64 篇视听生成论文，上一届是 32 篇。

ICML 2026 有多少篇视听生成论文？

ICML 2026 有 39 篇视听生成论文，上一届是 19 篇。

CVPR 2026 有多少篇视听生成论文？

CVPR 2026 有 149 篇视听生成论文，上一届是 113 篇。

哪个会议的视听生成论文最多？

CVPR，2023–2026 年共 383 篇。