零样本分类论文趋势 2023–2026

主要结论

• 零样本分类论文从 2023 年的 203 篇增长到 2026 年的 257 篇，三年间扩大到约 1.3 倍。
• CVPR 是发零样本分类论文最多的会议，2023–2026 年共 318 篇，占四个会议总量的 32%。
• 交叉方向里涨得最快的是强化学习：同时属于两个主题的论文从 2023 年的 7 篇增至 2026 年的 22 篇。
• 在「开放词汇视觉」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的视觉定位是 4.3 倍，零样本分类是 1.3 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 Minimal Interpretable Architecture for Zero-Shot Reconstruction of Dynamical Systems
• ICML 2026Beyond Rational Illusion: Behaviorally Realistic Strategic Classification
• ICLR 2026Boomerang Distillation Enables Zero-Shot Model Size Interpolation
• CVPR 2026AG-VAS Anchor-Guided Zero-Shot Visual Anomaly Segmentation with Large Multimodal Models
• NeurIPS 2025\(\varepsilon\)-Optimally Solving Two-Player Zero-Sum POSGs
• ICML 2025A Classification View on Meta Learning Bandits
• ICLR 2025Convergence and Implicit Bias of Gradient Descent on Continual Linear Classification
• CVPR 2025AA-CLIP: Enhancing Zero-Shot Anomaly Detection via Anomaly-Aware CLIP
• NeurIPS 2024BIOSCAN-5M: A Multimodal Dataset for Insect Biodiversity
• ICML 2024Agent Instructs Large Language Models to be General Zero-Shot Reasoners
• ICLR 2024AutoCast++: Enhancing World Event Prediction with Zero-shot Ranking-based Context Retrieval
• CVPR 2024Depth-aware Test-Time Training for Zero-shot Video Object Segmentation
• NeurIPS 2023Black-box Backdoor Defense via Zero-shot Image Purification
• ICML 2023A Simple Zero-shot Prompt Weighting Technique to Improve Prompt Ensembling in Text-Image Models
• ICLR 2023Benign Overfitting in Classification: Provably Counter Label Noise with Larger Models
• CVPR 2023(ML)$^2$P-Encoder: On Exploration of Channel-Class Correlation for Multi-Label Zero-Shot Learning
• NeurIPS 2026Align3D-AD: Cross-Modal Feature Alignment and Dual-Prompt Learning for Zero-shot 3D Anomaly Detection
• ICML 2026Calibrating Uncertainty for Zero-Shot Adversarial CLIP
• ICLR 2026Decoupling Primitive with Experts: Dynamic Feature Alignment for Compositional Zero-Shot Learning
• CVPR 2026DLVP-CLIP Enhancing Fine-Grained Zero-Shot Anomaly Detection via Dynamic Local Visual

全部零样本分类论文

共 192 篇，第 2 / 4 页

51. ICML 2026TextMesh4D: Zero-shot Text-to-4D Mesh Generation
52. ICML 2026Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancy
53. ICML 2026Treatment Responder Classification with Abstention
54. ICML 2026Universal Algorithm-Implicit Learning
55. ICML 2026URS: A Unified Neural Routing Solver for Cross-Problem Zero-Shot Generalization
56. ICML 2026VocSim A Training-free Benchmark for Zero-shot Content Identity in Single-source Audio
57. ICML 2026Von Mises-Fisher Mixture Model with Dynamic Shrinkage for Realistic Test-Time Transduction
58. ICML 2026Welfare-Optimal Classification with Accuracy Auctions
59. ICML 2026When Generalized Zero-Shot Learning Meets PU Learning: A Plug-and-Play Framework for Seen-Class Bias Mitigation
60. ICML 2026Zero-Flow Encoders
61. ICML 2026Zero-shot Active Mapping via Fused 360-BEV Representations and Vision–Language Models
62. ICML 2026Zero-Shot Off-Policy Learning
63. ICML 2026Zero-Shot Rankability: Revealing Latent Ordinal Structure in Multimodal Large Language Models via Language
64. ICML 2026ZeroDiff: Zero-Shot Time Series Reconstruction via Informed-Prior Diffusion
65. ICML 2025A Classification View on Meta Learning Bandits
66. ICML 2025A Generalization Theory for Zero-Shot Prediction
67. ICML 2025Alpha-SQL: Zero-Shot Text-to-SQL using Monte Carlo Tree Search
68. ICML 2025Bootstrapping Self-Improvement of Language Model Programs for Zero-Shot Schema Matching
69. ICML 2025CERTAIN: Context Uncertainty-aware One-Shot Adaptation for Context-based Offline Meta Reinforcement Learning
70. ICML 2025Conformity Score Averaging for Classification
71. ICML 2025DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning
72. ICML 2025Discovering a Zero (Zero-Vector Class of Machine Learning)
73. ICML 2025DMOSpeech: Direct Metric Optimization via Distilled Diffusion Model in Zero-Shot Speech Synthesis
74. ICML 2025Do Not Mimic My Voice : Speaker Identity Unlearning for Zero-Shot Text-to-Speech
75. ICML 2025EnsLoss: Stochastic Calibrated Loss Ensembles for Preventing Overfitting in Classification
76. ICML 2025ERICT: Enhancing Robustness by Identifying Concept Tokens in Zero-Shot Vision Language Models
77. ICML 2025Explanatory Instructions: Towards Unified Vision Tasks Understanding and Zero-shot Generalization
78. ICML 2025FIC-TSC: Learning Time Series Classification with Fisher Information Constraint
79. ICML 2025Fine-Grained Captioning of Long Videos through Scene Graph Consolidation
80. ICML 2025Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models
81. ICML 2025GenZSL: Generative Zero-Shot Learning Via Inductive Variational Autoencoder
82. ICML 2025Implicit Subgraph Neural Network
83. ICML 2025Improving Zero-Shot Adversarial Robustness in Vision-Language Models by Closed-form Alignment of Adversarial Path Simplices
84. ICML 2025Learning from True-False Labels via Multi-modal Prompt Retrieving
85. ICML 2025Learning Imbalanced Data with Beneficial Label Noise
86. ICML 2025Maximum Update Parametrization and Zero-Shot Hyperparameter Transfer for Fourier Neural Operators
87. ICML 2025Multi-band Frequency Reconstruction for Neural Psychoacoustic Coding
88. ICML 2025NegMerge: Sign-Consensual Weight Merging for Machine Unlearning
89. ICML 2025Neural Representational Consistency Emerges from Probabilistic Neural-Behavioral Representation Alignment
90. ICML 2025On the Provable Separation of Scales in Maximal Update Parameterization
91. ICML 2025One-Shot Heterogeneous Federated Learning with Local Model-Guided Diffusion Models
92. ICML 2025Online Linear Classification with Massart Noise
93. ICML 2025PAC-Bayes Analysis for Recalibration in Classification
94. ICML 2025Pessimism Principle Can Be Effective: Towards a Framework for Zero-Shot Transfer Reinforcement Learning
95. ICML 2025Proto Successor Measure: Representing the Behavior Space of an RL Agent
96. ICML 2025Provable Zero-Shot Generalization in Offline Reinforcement Learning
97. ICML 2025Provably Improving Generalization of Few-shot models with Synthetic Data
98. ICML 2025Random Registers for Cross-Domain Few-Shot Learning
99. ICML 2025Revisiting Neural Networks for Few-Shot Learning: A Zero-Cost NAS Perspective
100. ICML 2025Robust Automatic Modulation Classification with Fuzzy Regularization

常见问题

2023–2026 年四大会议一共有多少篇零样本分类论文？

共 992 篇，按年份分别是 2023 年 203 篇、2024 年 262 篇、2025 年 270 篇、2026 年 257 篇。

NeurIPS 2026 有多少篇零样本分类论文？

NeurIPS 2026 有 68 篇零样本分类论文，上一届是 86 篇。

ICLR 2026 有多少篇零样本分类论文？

ICLR 2026 有 44 篇零样本分类论文，上一届是 56 篇。

ICML 2026 有多少篇零样本分类论文？

ICML 2026 有 64 篇零样本分类论文，上一届是 49 篇。

CVPR 2026 有多少篇零样本分类论文？

CVPR 2026 有 81 篇零样本分类论文，上一届是 79 篇。

哪个会议的零样本分类论文最多？

CVPR，2023–2026 年共 318 篇。