理论 · optimization论文趋势 2023–2026

主要结论

• 理论 · optimization论文从 2023 年的 123 篇增长到 2026 年的 584 篇，三年间扩大到约 4.7 倍。
• NeurIPS 是发理论 · optimization论文最多的会议，2023–2026 年共 778 篇，占四个会议总量的 42%。
• 交叉方向里涨得最快的是低秩适配：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 31 篇。
• 在「理论与统计」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的理论 · learning theory是 5.6 倍，理论 · optimization是 4.7 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 2026A Mechanistic Investigation of Theory of Mind in a Large Language Model
• ICML 2026A Direct Second-Order Method for Solving Two-Player Zero-Sum Games
• ICLR 2026A Convergence Analysis of Adaptive Optimizers under Floating-point Quantization
• CVPR 2026TriSim Tri-Dimensional Similarity Modeling with Extreme Value Theory for False-Negative
• NeurIPS 20253BASiL: An Algorithmic Framework for Sparse plus Low-Rank Compression of LLMs
• ICML 2025Active Learning of Deep Neural Networks via Gradient-Free Cutting Planes
• ICLR 2025A Coefficient Makes SVRG Effective
• CVPR 2025Dynamic Stereotype Theory Induced Micro-expression Recognition with Oriented Deformation
• NeurIPS 2024A Combinatorial Algorithm for the Semi-Discrete Optimal Transport Problem
• ICML 2024A Statistical Theory of Regularization-Based Continual Learning
• ICLR 2024A Discretization Framework for Robust Contextual Stochastic Optimization
• NeurIPS 2023[Re] Bandit Theory and Thompson Sampling-guided Directed Evolution for Sequence Optimization
• ICML 2023Averaged Method of Multipliers for Bi-Level Optimization without Lower-Level Strong Convexity
• ICLR 2023Adaptive Optimization in the $\infty$-Width Limit
• CVPR 2023Deep Hashing With Minimal-Distance-Separated Hash Centers
• NeurIPS 2026Avoiding Obfuscation with Prover-Estimator Debate
• ICML 2026A General Framework for Dynamic Consistent Submodular Maximization
• ICLR 2026A Memory-Efficient Hierarchical Algorithm for Large-scale Optimal Transport Problems
• CVPR 2026FedAdamom Adaptive Momentum for Improved Generalization in Federated Optimization
• NeurIPS 2025A Beyond-Worst-Case Analysis of Greedy k-means++

全部理论 · optimization论文

共 609 篇，第 3 / 13 页

101. ICLR 2026Graph-based Nearest Neighbors with Dynamic Updates via Random Walks
102. ICLR 2026Group Representational Position Encoding
103. ICLR 2026Group-Normalized Implicit Value Optimization for Language Models
104. ICLR 2026Grouping Nodes with known Value Differences: A lossless UCT-based Abstraction Algorithm
105. ICLR 2026Guaranteed Simply Connected Mesh Reconstruction from an Unorganized Point Cloud
106. ICLR 2026Half-order Fine-Tuning for Diffusion Model: A Recursive Likelihood Ratio Optimizer
107. ICLR 2026Harmonized Cone for Feasible and Non-conflict Directions in Training Physics-Informed Neural Networks
108. ICLR 2026HiFo-Prompt: Prompting with Hindsight and Foresight for LLM-based Automatic Heuristic Design
109. ICLR 2026High Probability Bounds for Non-Convex Stochastic Optimization with Momentum
110. ICLR 2026High-Probability Bounds for the Last Iterate of Clipped SGD
111. ICLR 2026Hinge Regression Tree: A Newton Method for Oblique Regression Tree Splitting
112. ICLR 2026How does the optimizer implicitly bias the model merging loss landscape?
113. ICLR 2026How hard is learning to cut? Trade-offs and sample complexity
114. ICLR 2026Hyperbolic Aware Minimization: Implicit Bias for Sparsity
115. ICLR 2026IDEAL: Data Equilibrium Adaptation for Multi-Capability Language Model Alignment
116. ICLR 2026Implicit Bias and Loss of Plasticity in Matrix Completion: Depth Promotes Low-Rankness
117. ICLR 2026Implicit Bias of Per-sample Adam on Separable Data: Departure from the Full-batch Regime
118. ICLR 2026Implicit bias produces neural scaling laws in learning curves, from perceptrons to deep networks
119. ICLR 2026Implicit Inversion turns CLIP into a Decoder
120. ICLR 2026Improved $\ell_{p}$ Regression via Iteratively Reweighted Least Squares
121. ICLR 2026Improved high-dimensional estimation with Langevin dynamics and stochastic weight averaging
122. ICLR 2026Improving Feasibility via Fast Autoencoder-Based Projections
123. ICLR 2026Improving LLM-based Global Optimization with Search Space Partitioning
124. ICLR 2026Improving Online-to-Nonconvex Conversion for Smooth Optimization via Double Optimism
125. ICLR 2026Infinite Horizon Markov Economies
126. ICLR 2026INSTANT: Compressing Gradients and Activations for Resource-Efficient Training
127. ICLR 2026Interactive Learning of Single-Index Models via Stochastic Gradient Descent
128. ICLR 2026Landing with the Score: Riemannian Optimization through Denoising
129. ICLR 2026Language Identification in the Limit with Computational Trace
130. ICLR 2026Larger Datasets Can Be Repeated More: A Theoretical Analysis of Multi-Epoch Scaling in Linear Regression
131. ICLR 2026Learning a Game by Paying the Agents
132. ICLR 2026Learning Admissible Heuristics for A*: Theory and Practice
133. ICLR 2026Learning to Adapt: In-Context Learning Beyond Stationarity
134. ICLR 2026Learning to Answer from Correct Demonstrations
135. ICLR 2026Learning to Recall with Transformers Beyond Orthogonal Embeddings
136. ICLR 2026Learning to Segment for Vehicle Routing Problems
137. ICLR 2026Learning to Solve Orienteering Problem with Time Windows and Variable Profits
138. ICLR 2026LEGACY: A Lightweight Dynamic Gradient Compression Strategy for Distributed Deep Learning
139. ICLR 2026Lifelong Learning with Behavior Consolidation for Vehicle Routing
140. ICLR 2026Lightweight Transformer for EEG Classification via Balanced Signed Graph Algorithm Unrolling
141. ICLR 2026LMask: Learn to Solve Constrained Routing Problems with Lazy Masking
142. ICLR 2026LogART: Pushing the Limit of Efficient Logarithmic Post-Training Quantization
143. ICLR 2026LoRA-S: An Efficient Low Rank Adaptation scheme via Sylvester equation
144. ICLR 2026Machine Unlearning under Retain–Forget Entanglement
145. ICLR 2026MaskCO: Masked Generation Drives Effective Representation Learning and Exploiting for Combinatorial Optimization
146. ICLR 2026MILPnet: A Multi-Scale Architecture with Geometric Feature Sequence Representations for Advancing MILP Problems
147. ICLR 2026Minor First, Major Last: A Depth-Induced Implicit Bias of Sharpness-Aware Minimization
148. ICLR 2026Monotone Near-Zero-Sum Games: A Generalization of Convex-Concave Minimax
149. ICLR 2026MT-DAO: Multi-Timescale Distributed Adaptive Optimizers with Local Updates
150. ICLR 2026Multimodal LLM-assisted Evolutionary Search for Programmatic Control Policies

常见问题

2023–2026 年四大会议一共有多少篇理论 · optimization论文？

共 1,871 篇，按年份分别是 2023 年 123 篇、2024 年 494 篇、2025 年 670 篇、2026 年 584 篇。

NeurIPS 2026 有多少篇理论 · optimization论文？

NeurIPS 2026 有 119 篇理论 · optimization论文，上一届是 318 篇。

ICLR 2026 有多少篇理论 · optimization论文？

ICLR 2026 有 244 篇理论 · optimization论文，上一届是 196 篇。

ICML 2026 有多少篇理论 · optimization论文？

ICML 2026 有 214 篇理论 · optimization论文，上一届是 147 篇。

CVPR 2026 有多少篇理论 · optimization论文？

CVPR 2026 有 7 篇理论 · optimization论文，上一届是 9 篇。

哪个会议的理论 · optimization论文最多？

NeurIPS，2023–2026 年共 778 篇。