The Future of Artificial Intelligence : The Next Step – Part 1 of 2
If you'd like to make a donation to help me get the message out and support my efforts, you can PayPal or Patreon me at these links:
https://paypal.me/wai69
https://www.patreon.com/waihtsang
This talk is about biologically inspired Artificial Intelligence and in particular the application of the ideas behind the Fractal Brain Theory in relation to the creation of a new generation of AI and Machine Learning technologies.
In recent years the limitations of existing AI and machine learning methods, have become increasingly apparent. So at the same time there has emerged the need for a better, more effective and also a more energy and compute efficient way of doing things.
It is more widely believed nowadays that it is through applying ideas derived from learning about biological intelligence and how real brains work, that the goal of creating artificial brains and intelligence will finally be achieved. We'll be demonstrating that this is absolutely the case.
The way existing AI 'learns' also leaves a lot to be desired. In particular it is desired that AI should learn like humans and biological systems, i.e. which learn 'on the fly', with a few shots or iterations vs many thousands, in a way that transfers previous learning without erasing it altogether, and in a way that is unsupervised, incremental and compositional. We'll explain how this can be done, using biology as our guide.
There are a few ideas that are floating around, which are believed to be key to creating better AI. For instance, if we could reconcile the symbolic and sub-symbolic levels of processing, if we had a good way of representing space and time hierarchically, if we could incorporate predictive, causal and generative mechanisms into our attempts at AI; then perhaps we would have something closer to true AI.
If you’d like to make a donation to help me get the message out and support my efforts, you can PayPal or Patreon me at these links:
https://paypal.me/wai69
https://www.patreon.com/waihtsang
This talk is about biologically inspired Artificial Intelligence and in particular the application of the ideas behind the Fractal Brain Theory in relation to the creation of a new generation of AI and Machine Learning technologies.
In recent years the limitations of existing AI and machine learning methods, have become increasingly apparent. So at the same time there has emerged the need for a better, more effective and also a more energy and compute efficient way of doing things.
It is more widely believed nowadays that it is through applying ideas derived from learning about biological intelligence and how real brains work, that the goal of creating artificial brains and intelligence will finally be achieved. We’ll be demonstrating that this is absolutely the case.
The way existing AI ‘learns’ also leaves a lot to be desired. In particular it is desired that AI should learn like humans and biological systems, i.e. which learn ‘on the fly’, with a few shots or iterations vs many thousands, in a way that transfers previous learning without erasing it altogether, and in a way that is unsupervised, incremental and compositional. We’ll explain how this can be done, using biology as our guide.
There are a few ideas that are floating around, which are believed to be key to creating better AI. For instance, if we could reconcile the symbolic and sub-symbolic levels of processing, if we had a good way of representing space and time hierarchically, if we could incorporate predictive, causal and generative mechanisms into our attempts at AI; then perhaps we would have something closer to true AI.
