CORNN凸优化递归神经网络方法论。将RNN训练转化为凸优化问题，训练速度比传统方法快100倍，支持百万参数RNN在标准计算机上亚分钟级训练。适用于大规模神经记录实时建模、神经动力学推断、吸引子结构恢复。触发词：RNN训练、凸优化、神经动力学、实时建模、数据约束RNN、convex optimization、recurrent neural network、neural dynamics inference。
no risky patterns detected
Type: Agent skill (SKILL.md) · Category: LLM & Model APIs · Source: https://github.com/hiyenwong/ai_collection/tree/HEAD/collection/skills/ai-ml/cor
| Capabilities (fewer powerful permissions = higher) | 50/50 |
| Maintenance | 21/35 |
| Documentation | 15/15 |
Always review files before installing. How scores work. Request a correction