So just to understand, my sensory motor inputs tell my neocortex "how" I'm moving (index finger up, middle finger down, etc.) and the output of that sensory sequence is compared to the "what" I'm experiencing (i.e.) it's features then understood by relating the context of the two?
+Benjamin Jordan sensory features are proximal input to a layer, the location on the object where the sensation was felt is the distal input. The input layer combines these into another representation contained both data. The “what” is another matter we will talk more about (active research topic).
In the future, professor Hawkins WILL be recognized as the creator/inventor/founder of machine intelligence! In our AI age, his model is the ONLY correct one.
What happens with a HTM unit behaviour for a specified goal in the environment when the unit is reconnected to another sensor, in case when the new sensor has the same interface as the previous one and in case when the interface is different?
It might take a long time, but eventually the system would adapt to the new input. It's behavior would change, but hard to say how without specific environments.
Your background music at the beginning is very annoying and totally unnecessary. Apart from that, I believe Numeta's HTM approach is the correct path to general artificial intelligence.
There is no reason we cannot understand ourselves. You are applying an artificial restriction based on ideology. We are looking for how it really works. The brain is just another scientific mystery to solve. We also used to think the world was flat and it was foolish to sail out too far. In our case, we don't even have to put ourselves at physical risk to research these discoveries. :)
![Jeff Hawkins - Human Brain Project Keynote [Screencast]](http://i.ytimg.com/vi/upXsle3yB6A/mqdefault.jpg)
![Jeff Hawkins - Human Brain Project Keynote [Screencast]](/img/tr.png)
Been waiting for this for years! *EDIT: the new sensorimotor theory/mechanism.
so good! thanks for the awesome audio quality too!
So just to understand, my sensory motor inputs tell my neocortex "how" I'm moving (index finger up, middle finger down, etc.) and the output of that sensory sequence is compared to the "what" I'm experiencing (i.e.) it's features then understood by relating the context of the two?
+Benjamin Jordan sensory features are proximal input to a layer, the location on the object where the sensation was felt is the distal input. The input layer combines these into another representation contained both data. The “what” is another matter we will talk more about (active research topic).
HTM School very interesting, thanks!
In the future, professor Hawkins WILL be recognized as the creator/inventor/founder of machine intelligence! In our AI age, his model is the ONLY correct one.
What happens with a HTM unit behaviour for a specified goal in the environment when the unit is reconnected to another sensor, in case when the new sensor has the same interface as the previous one and in case when the interface is different?
It might take a long time, but eventually the system would adapt to the new input. It's behavior would change, but hard to say how without specific environments.
What does motor sensory inference give? What can it do other than previous model?
Imagine a sensor has the ability to move. SMI allows the TM algorithm to store allocentric object representations for retrieval.
Your background music at the beginning is very annoying and totally unnecessary. Apart from that, I believe Numeta's HTM approach is the correct path to general artificial intelligence.
You are trying to interpret brain in brain, so, ultimately you can not understand brain.
There is no reason we cannot understand ourselves. You are applying an artificial restriction based on ideology. We are looking for how it really works. The brain is just another scientific mystery to solve. We also used to think the world was flat and it was foolish to sail out too far. In our case, we don't even have to put ourselves at physical risk to research these discoveries. :)