数学界无视「30年漏洞」，GPT-5一眼看穿，陶哲轩：AI科研革命开始了
"优化" 相关文章
Fast PINN Eigensolvers via Biconvex Reformulation
Reasoning Planning for Language Models
A Dual Large Language Models Architecture with Herald Guided Prompts for Parallel Fine Grained Traffic Signal Control
Benchmarking Generative AI Against Bayesian Optimization for Constrained Multi-Objective Inverse Design
[Web Dev] 为什么没有精简 WebKit？
[Web Dev] 为什么没有精简 WebKit？
Adaptive Data Flywheel: Applying MAPE Control Loops to AI Agent Improvement
[程序员] 微信的聊天记录管理逻辑是不是有点问题
Defeating the Training-Inference Mismatch via FP16
Distributional Multi-objective Black-box Optimization for Diffusion-model Inference-time Multi-Target Generation
The Neural Differential Manifold: An Architecture with Explicit Geometric Structure
Scheduling Your LLM Reinforcement Learning with Reasoning Trees
[输入法] 需要一个尊重上下文的输入法
参数空间对称性：深度学习理论的统一几何框架
NeurIPS 2025 | 利用物理信息神经网络加速大规模变分量子算法
Design and Optimization of Cloud Native Homomorphic Encryption Workflows for Privacy-Preserving ML Inference
Investigating Intra-Abstraction Policies For Non-exact Abstraction Algorithms
Softmax is $1/2$-Lipschitz: A tight bound across all $\ell_p$ norms
抖音云游戏调度优化实践
