arXiv · Neural and Evolutionary Computing· Johnson Zhou, Daniel Tanneberg, Forough Habibollahi, Alon Loeffler, Kiaran Lawson, Valentina Baccetti, Kwaku Dad Abu-Bonsrah, Candice Desouza, Finn Doensen, Bradley Watmuff, Daria Kornienko, Azin Azadi, Justin Leigh Bourke, Bernhard Sendhoff, Brett J. Kagan·· 16 小时前AI 评分31
Embodied Neurocomputation：生物神经培养物接口框架与规模化任务驱动验证
Embodied Neurocomputation: A Framework for Interfacing Biological Neural Cultures with Scaled Task-Driven Validation
AI 导读
研究者提出 Embodied Neurocomputation 框架，用于系统级优化传统硅计算接口与生物神经网络（BNN）之间的编码与解码机制。团队对在模拟网格世界中沿气味梯度做闭环导航的 BNN 智能体完成首次大规模编码参数优化，评估约 1,300 个参数组合、超过 4,000 小时实时智能体—环境交互，筛出 12 个在多轮次中持续表现出学习的配置。
来源：arXiv · Neural and Evolutionary Computing · arxiv.org