Current AGI proposals are static ways; Movement and interaction with external world is the key fundation for AGI development. When living in a real world, we cotinuously get quite different infoamtion of objects ,we gradually get understanding of the world from what we see and feel, things that changed or unchanged. We gradually konw what will change and what wiil not change due to some factor's variations. We gradually know the law of physics and the characteristics of human beings. When we see,touch and feel the real world, our brain might build a tree or network description of the world almost simutaenously, some of the node of the network description might be important and others are not. Through competition and cross verification of different sensing organs, some nodes and the way to build those nodes gradually win the competition, and we developed a reasonale way to understand the enviroment.
JZTANG ZJU From: [email protected] Date: 2019-05-29 16:53 To: AGI Subject: Re: Re: [agi] Do Neural Networks Need To Think Like Humans? I think your way is a minor step toward right direction. However , it should be totally unsupervised. Knowledge should be gradually developed though intraction with enviroment. JZTANG ZJU From: keghnfeem Date: 2019-05-28 21:28 To: AGI Subject: Re: [agi] Do Neural Networks Need To Think Like Humans? @jztang i am working on this. It is a unsupervised RNN GAN. i call it un-Gan. Vicarious has part of it which is a supervised version: https://www.vicarious.com/2017/10/26/common-sense-cortex-and-captcha/ Artificial General Intelligence List / AGI / see discussions + participants + delivery options Permalink ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T4aa81fd9912dbd39-Mdf1ce592bcb4aac76d77d3eb Delivery options: https://agi.topicbox.com/groups/agi/subscription
