Tweet

Carlos E. Perez

14 Nov, 6 tweets, 3 min read

@blamlab

Yesterday's Learning Salon with Gary Marcus. The last 30 minutes were excellent (after the guest left). The best conclusion: @blamlab AI is the subversive idea that cognitive psychology can be formalized.
crowdcast.io/e/learningsalo…

Important to realize that a description of a missing cognitive functionality does not have enough precision or leave enough hints on how this is implemented in the brain. Implementations in code do not imply how it is implemented in the brain.

Another distinction that is important that there is a disagreement on how to do research. The Deep Learning community has argued that we should not constrain ourselves with a-priori hypothesis that may be wrong. Let the learning system discover the algorithms.

This differs from the more conventional AI approach where we describe the functionality and seek to formulate algorithms to implement these. However a combination of learning systems and hypothesis-making about learning systems is the path to progress.

@garymarcus

One important nugget of wisdom from @garymarcus is where he notes the importance of intention and how the brain arrives at the procedure to achieve the intention. Brian Cantwell Smith is correct when he says that computation is the interplay of intention and mechanism.

@blamlab

I just realized I missed 2 previous episodes. Time to watch them. The best parts are the discussions. Thanks to @blamlab @neuro_data @criticalneuro for an extremely valuable forum that bridges ideas in neuroscience, AI and psychology.

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @IntuitMachine

Carlos E. Perez

@IntuitMachine

15 Nov

Analysis of QAnon by a game designer. Everyone should read! medium.com/curiouserinsti…

QAnon method is like the movie Inception on a mass scale. Planting seeds of misinformation so that its victims generate their understanding of alternative reality for themselves.

The author concludes that this isn't a movement that grew organically, but rather one that is orchestrated with big money.

Read 24 tweets

Carlos E. Perez

@IntuitMachine

14 Nov

Jay McClelland on What's missing in Deep Learning crowdcast.io/e/learningsalo…

He argues against innate systematic generalization in humans and it is something that we acquire.

Thus he argues that to achieve systematic generalization we need to devise machines that learn how to do systematic generalization. That is, a meta-solution to the problem.

Read 21 tweets

Carlos E. Perez

@IntuitMachine

13 Nov

The more try to understand cognition, the more you realize how long the journey may be required to get human-like general intelligence.

Our frameworks of understanding cognition are getting better. However, one has to understand that cognition arises through emergence in complex adaptive systems. These systems are very difficult to set up and replicate.

To get an intuition of how large the gap truly is, one only needs to observe how awkward and non-organic in our present-day robots. Why can't they perform with the nimbleness of honey bees?

Read 5 tweets

Carlos E. Perez

@IntuitMachine

13 Nov

The idealization of the ethnic peasantry as the one true national class is the generating condition that lead to genocides in Nazi Germany, Armenia, and Cambodia. It is fueled by the resentment of the elite as the root of their own misery.

We need to learn from history and ask why a country like Cambodia will put a quarter of its population to death only because they were experts in different crafts. en.wikipedia.org/wiki/Cambodian…

What collectively drives people to kill people on a mass scale? What makes people ignore their natural empathy for others? It is the collective delusion that the existence of another is the reason for one's misery.

Read 19 tweets

Carlos E. Perez

@IntuitMachine

12 Nov

Our brains see affordances, not because we see 3D objects, but rather how we move to see our 3D worlds. This is different from inverse graphics.

Inverse graphics implies that a representation of a 3D object is created in our minds. If we assume the 'lazy brain hypothesis' then this doesn't make sense because it is wasteful computation.

When we look at a Rubik's cube, we don't instantly know where which colors are on which sides. We are aware of the layout only when we attend to it. It is an active form of perception rather than a photographic snapshot.

Read 5 tweets

Carlos E. Perez

@IntuitMachine

5 Nov

Modularity is critically important for AGI, but we should avoid a naive formulation of modularity.

The most developed notion of modularity comes from computer science. We have notions of encapsulation and all kinds of composition design patterns that are strategically employed to trade-off one concern over another.

Modularity thus is understood as a controlled coordination mechanism between interacting parts. Certain information is allowed to be malleable while other kinds must remain immutable.

Read 10 tweets

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!