cs.AI updates on arXiv.org 07月08日 13:54
Neural Inhibition Improves Dynamic Routing and Mixture of Experts
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

本文提出在神经网络中引入神经抑制机制,以优化动态路由模型,提高其性能。实验证明,此方法能有效提升模型在通用任务上的表现,并呼吁更多研究投入。

arXiv:2507.03221v1 Announce Type: cross Abstract: To be effective, efficient, and diverse, deep learning models need to dynamically choose its architecture based on signals from a population of neurons. We hypothesize dynamic routing models can be improved with neural inhibition in those neural populations. This means signals commonly shared among the various modes of data statistics can be inhibited so that the routing model can choose a specialized expert path for each data sample. Only through inhibition is the routing mechanism able to effectively select neural pathways. We believe this is an under-studied and under-verified implementation methodology for Mixture-of-Experts, dynamic routing, and transformer language models. We provide experimental evidence that the neural inhibition algorithm significantly boosts the performance of general tasks and motivates more effort to be invested in this research direction.

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

神经抑制 动态路由 深度学习 模型性能 Mixture-of-Experts
相关文章