通用机器学习 · evaluation论文趋势 2023–2026

主要结论

• 通用机器学习 · evaluation论文从 2023 年的 53 篇增长到 2026 年的 230 篇，三年间扩大到约 4.3 倍。
• NeurIPS 是发通用机器学习 · evaluation论文最多的会议，2023–2026 年共 270 篇，占四个会议总量的 45%。
• 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 47 篇。
• 在「通用机器学习」的 6 个主题里，它的 2026 年论文数排第 4；同组增长最快的通用机器学习 · reinforcement learning是 2.0 倍，通用机器学习 · evaluation是 4.3 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 2026AI Construct Lexis: An Ontology of the Hidden Assumptions in AI Evaluation
• ICML 2026A Computational Framework for Evaluating Human-likeness in LLMs' Open-ended Human Behaviors
• ICLR 2026Unlearning Evaluation through Subset Statistical Independence
• CVPR 2026Bridging the Perception Gap in Image Super-Resolution Evaluation
• NeurIPS 2025Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search
• ICML 2025AKRMap: Adaptive Kernel Regression for Trustworthy Visualization of Cross-Modal Embeddings
• ICLR 2025On Speeding Up Language Model Evaluation
• CVPR 2025ODE: Open-Set Evaluation of Hallucinations in Multimodal Large Language Models
• NeurIPS 2024Accuracy is Not All You Need
• ICML 2024A Fine-grained Analysis of Fitted Q-evaluation: Beyond Parametric Models
• ICLR 2024BatchPrompt: Accomplish more with less
• CVPR 2024Animating General Image with Large Visual Motion Model
• NeurIPS 2023AQuA: A Benchmarking Tool for Label Quality Assessment
• ICML 2023Regions of Reliability in the Evaluation of Multivariate Probabilistic Forecasts
• ICLR 2023Beyond calibration: estimating the grouping loss of modern neural networks
• CVPR 2023GeneCIS: A Benchmark for General Conditional Image Similarity
• NeurIPS 2026AMUSE: Anytime Muon with Stable Gradient Evaluation
• ICML 2026A Two-Tier Perspective on Inference-Time Parallelism in Multi-Agent LLM Systems
• ICLR 2026Choices Speak Louder than Questions
• CVPR 2026Towards Reliable Evaluation of Adversarial Robustness for Spiking Neural Networks

全部通用机器学习 · evaluation论文

共 208 篇，第 2 / 5 页

51. ICML 2025MCU: An Evaluation Framework for Open-Ended Game Agents
52. ICML 2025Measuring Diversity in Synthetic Datasets
53. ICML 2025MetricEmbedding: Accelerate Metric Nearness by Tropical Inner Product
54. ICML 2025Minerva: A Programmable Memory Test Benchmark for Language Models
55. ICML 2025Modified K-means Algorithm with Local Optimality Guarantees
56. ICML 2025Multiple-policy Evaluation via Density Estimation
57. ICML 2025Off-Policy Evaluation under Nonignorable Missing Data
58. ICML 2025On the Diversity of Adversarial Ensemble Learning
59. ICML 2025Overestimation in LLM Evaluation: A Controlled Large-Scale Study on Data Contamination’s Impact on Machine Translation
60. ICML 2025PaperBench: Evaluating AI’s Ability to Replicate AI Research
61. ICML 2025Position: AI Competitions Provide the Gold Standard for Empirical Rigor in GenAI Evaluation
62. ICML 2025Position: AI Evaluation Should Learn from How We Test Humans
63. ICML 2025Position: In-House Evaluation Is Not Enough. Towards Robust Third-Party Evaluation and Flaw Disclosure for General-Purpose AI
64. ICML 2025Position: Stop treating `AGI' as the north-star goal of AI research
65. ICML 2025Prompt-to-Leaderboard: Prompt-Adaptive LLM Evaluations
66. ICML 2025Provable Efficiency of Guidance in Diffusion Models for General Data Distribution
67. ICML 2025RAGGED: Towards Informed Design of Scalable and Stable RAG Systems
68. ICML 2025Raising the Bar: Investigating the Values of Large Language Models via Generative Evolving Testing
69. ICML 2025RBench: Graduate-level Multi-disciplinary Benchmarks for LLM & MLLM Complex Reasoning Evaluation
70. ICML 2025RE-IMAGINE: Symbolic Benchmark Synthesis for Reasoning Evaluation
71. ICML 2025Reflection-Bench: Evaluating Epistemic Agency in Large Language Models
72. ICML 2025Regression for the Mean: Auto-Evaluation and Inference with Few Labels through Post-hoc Regression
73. ICML 2025Rethinking Causal Ranking: A Balanced Perspective on Uplift Model Evaluation
74. ICML 2025Robust ML Auditing using Prior Knowledge
75. ICML 2025SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?
76. ICML 2025SyncMind: Measuring Agent Out-of-Sync Recovery in Collaborative Software Engineering
77. ICML 2025The Berkeley Function Calling Leaderboard (BFCL): From Tool Use to Agentic Evaluation of Large Language Models
78. ICML 2025The Hidden Joules: Evaluating the Energy Consumption of Vision Backbones for Progress Towards More Efficient Model Inference
79. ICML 2025To Each Metric Its Decoding: Post-Hoc Optimal Decision Rules of Probabilistic Hierarchical Classifiers
80. ICML 2025Triple-Optimistic Learning for Stochastic Contextual Bandits with General Constraints
81. ICML 2025TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories
82. ICML 2025Uncertainty Quantification for LLM-Based Survey Simulations
83. ICML 2025Unified K-Means Clustering with Label-Guided Manifold Learning
84. ICML 2025WATCH: Adaptive Monitoring for AI Deployments via Weighted-Conformal Martingales
85. NeurIPS 2025A Finite Sample Analysis of Distributional TD Learning with Linear Function Approximation
86. NeurIPS 2025A General-Purpose Theorem for High-Probability Bounds of Stochastic Approximation with Polyak Averaging
87. NeurIPS 2025A Principled Path to Fitted Distributional Evaluation
88. NeurIPS 2025Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search
89. NeurIPS 2025AGI-Elo: How Far Are We From Mastering A Task?
90. NeurIPS 2025AI Debate Aids Assessment of Controversial Claims
91. NeurIPS 2025AI Testing Should Account for Sophisticated Strategic Behaviour
92. NeurIPS 2025Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs
93. NeurIPS 2025AnaCP: Toward Upper-Bound Continual Learning via Analytic Contrastive Projection
94. NeurIPS 2025APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplay
95. NeurIPS 2025Are Large Reasoning Models Good Translation Evaluators? Analysis and Performance Boost
96. NeurIPS 2025Ascent Fails to Forget
97. NeurIPS 2025Atomic Thinking of LLMs: Decoupling and Exploring Mathematical Reasoning Abilities
98. NeurIPS 2025Attention! Your Vision Language Model Could Be Maliciously Manipulated
99. NeurIPS 2025Automated Detection of Visual Attribute Reliance with a Self-Reflective Agent
100. NeurIPS 2025Benchmarking End-To-End Performance of AI-Based Chip Placement Algorithms

常见问题

2023–2026 年四大会议一共有多少篇通用机器学习 · evaluation论文？

共 606 篇，按年份分别是 2023 年 53 篇、2024 年 115 篇、2025 年 208 篇、2026 年 230 篇。

NeurIPS 2026 有多少篇通用机器学习 · evaluation论文？

NeurIPS 2026 有 57 篇通用机器学习 · evaluation论文，上一届是 124 篇。

ICLR 2026 有多少篇通用机器学习 · evaluation论文？

ICLR 2026 有 19 篇通用机器学习 · evaluation论文，上一届是 19 篇。

ICML 2026 有多少篇通用机器学习 · evaluation论文？

ICML 2026 有 148 篇通用机器学习 · evaluation论文，上一届是 58 篇。

CVPR 2026 有多少篇通用机器学习 · evaluation论文？

CVPR 2026 有 6 篇通用机器学习 · evaluation论文，上一届是 7 篇。

哪个会议的通用机器学习 · evaluation论文最多？

NeurIPS，2023–2026 年共 270 篇。