Update vision-of-the-thousand-brains-project.md #96
Conversation
minor tweaks
|
Thank you for your contribution @RichMorin! It appears that you haven't signed our Contributor License Agreement yet. Please visit this link and sign. Note New signatures are processed during the work week. It may take some time before your signature is processed. You will be invited to the Numenta |
|
Hi @RichMorin, I see a signed CLA with |
|
I'm afraid I don't understand what you are asking me to do. More generally, is there a conflict between the names RichMorin and Rich_Morin? ELI5...
…-r
On Dec 10, 2024, at 09:59, Tristan Slominski ***@***.***> wrote:
Hi @RichMorin, I see a signed CLA with Rich_Morin GitHub username. Please use RichMorin to pass the CLA check.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you were mentioned.
|
|
Hi @RichMorin. Thank you for signing the CLA. However, the The GitHub username on the CLA: For the CLA check to pass, the GitHub username on record in the CLA form should match the GitHub username used for pull requests. If you would please visit this link and sign ensuring that the form entry for |
tristanls
added
documentation
|
Thank you for signing the CLA! @codeallthethingz, I'm tagging you to review these changes to high-level goal language. FYI @vkakerbeck. |
|
I'm going to swap in @vkakerbeck as the reviewer for this page as it is the vision page that I transcribed from a paper she authored. |
There was a problem hiding this comment.
Thanks for the updates! Its a bit hard to see the diffs here so I used https://www.diffchecker.com/text-compare/ but even with that the last paragraph didn't show anything. Lmk if I am missing something.
|
|
||
| # Reference Frames | ||
|
|
||
| A second differentiator is that our sensorimotor systems learn structured models, using _reference frames_, coordinate systems within which locations and rotations can be represented. The models keep track of where their sensors are relative to things in the world. They are learned by assigning sensory observations to locations in reference frames. In this way, the models learned by sensorimotor systems are structured, similar to CAD models in a computer. This allows the system to quickly learn the structure of the world and how to manipulate objects to achieve a variety of goals, what is sometimes referred to as a 'world model'. As with sensorimotor learning, reference frames are used throughout all levels of information processing, including the representations of not only environments, but also physical objects and abstract concepts - even the simplest representations in the proposed architecture are represented within a reference frame. | ||
|
|
||
| # Human-like Learning | ||
|
|
||
| There are numerous advantages to sensorimotor learning and reference frames. At a high level, you can think about all the ways humans are different from today's AI. We learn quickly and continuously, constantly updating our knowledge of the world as we go about our day. We do not have to undergo a lengthy and expensive training phase to learn something new. We interact with the world and manipulate tools and objects in sophisticated ways that leverage our knowledge of how things are structured. For example, we can explore a new app on our phone and quickly figure out what it does and how it works based on other apps we know. We actively test hypotheses to fill in the gaps in our knowledge. We also learn from multiple sensors and our different sensors work together seamlessly. For example, we may learn what a new tool looks like with a few glances and then immediately know how to grab and interact with the object via touch. | ||
| There are numerous advantages to sensorimotor learning and reference frames. At a high level, you can think about all the ways humans are different from today's AI. We learn quickly and continuously, constantly updating our knowledge of the world as we go about our day. We do not have to undergo a lengthy and expensive training phase to learn something new. We interact with the world and manipulate tools and objects in sophisticated ways that leverage our knowledge of how things are structured. For example, we can explore a new app on our phone and quickly figure out what it does and how it works based on other apps we know. We actively test hypotheses to fill in the gaps in our knowledge. We also learn from multiple sensors and our different sensors work together seamlessly. For example, we may learn what a new tool looks like with a few glances and then immediately know how to grab and interact with the object via touch. Finally, the basis for decisions is considerably less opaque than that found in Large Language Models, etc. |
There was a problem hiding this comment.
I wouldn't end the last sentence of the paragraph with ",etc." How about something like "Overall, our decisions and interactions with the world are based on structured models that we can leverage and combine in novel way to adapt to the wide range of tasks and circumstances we face every day."
We could add something about LLMs being just general function approximations, learning the statistical regularities in a dataset but I think that would go into too much depth for a vision page.
| One of the most important discoveries about the brain is that most of what we think of as intelligence, from seeing, to touching, to hearing, to conceptual thinking, to language, is created by a common neural algorithm. All aspects of intelligence are created by the same sensorimotor mechanism. In the neocortex, this mechanism is implemented in each of the thousands of cortical columns. This means we can create many different types of intelligent systems using a set of common building blocks. The architecture we are creating is built on this premise. Monty will provide the core components and developers will then be able to assemble widely varying AI and robotics applications using these components in different numbers and arrangements. Any engineer will be able to create AI applications using the Platform without requiring huge computational resources or background knowledge. |
There was a problem hiding this comment.
I don't see any difference here and also diff checker doesn't show anything. Am I missing something? If not, can you remove this "change" from the PR?
There was a problem hiding this comment.
The last paragraph adds a new line to the end of the file. These don't show up in GitHub.
There was a problem hiding this comment.
Ahh I see :D You would think github could show this as a bit of a smaller diff than the entire paragraph... In that case, yes let's keep the change in this PR.
There was a problem hiding this comment.
I couldn't find it anywhere on GitHub, but I wanted to see if development mode showed it. Turns out that it does. Press the . key and enter the development mode (whatever it is called), and you'll see the newline.
Screen.Recording.2024-12-12.at.09.47.48.mov
minor tweaks