On Intelligence

Sandra Blakeslee and Jeff Hawkins

Summary: This book had an impact on me while working on my PhD thesis as it resonated with my view of artificial (machine) intelligence.

Score: 85 / 100

This book had an impact on me while working on my PhD thesis as it resonated with my view of artificial (machine) intelligence.

I think that I borrowed it from NLP centre’s library and it is a book from Pavel Rychlý.

Lex Fridman’s podcast with Jeff

• Criterions for brain models
  • inclusion of time in brain function
  • importance of feedback
  • model should account for neocortex structure
  • auto-associative memory
  • retrieving patterns (p. 30) with mesy/noisy pattern input
  • ibid. feedback with delay: storing sequences
  • p. 32: "Our intuition is often the biggest obstacle to discovering the truth"
  • p. 34: "The best solutions to scientific problems are simple and elegant."
• Spatial vs. temporal robustness
  • invariant form of memory vs. invariant treatmen / processing of assoc. memory (my note)
  • what is stored doesnt need to be invariant: an algorithm be variant-proof… (my note)
• příklad s predikcí noty v písni: jestli mám v hlavě invariantní reprezentaci písně (intervaly), tak jaktože odhadnu přesně výšku tónu? vždyť Jeff psal, že vstup se vždycky mění na invariantní reprezentaci == chyba jeho úvahy?

Vernon Mountcastle, a neuroscientist at Johns Hopkins University in Baltimore. In 1978 he published a paper titled “An Organizing Principle for Cerebral Function.” (Thursday, January 08, 2015, 12:10 PM, page 735-36)

Mountcastle was right. There is a single powerful algorithm implemented by every region of cortex. If you connect regions of cortex together in a suitable hierarchy and provide a stream of input, it will learn about its environment. Therefore, there is no reason for intelligent machines of the future to have the same senses or capabilities as we humans. The cortical algorithm can be deployed in novel ways, with novel senses, in a machined cortical sheet so that genuine, flexible intelligence emerges outside of biological brains. (Thursday, January 08, 2015, 07:14 PM, page 810-13)

Vision is more like a song than a painting. (Thursday, January 08, 2015, 07:19 PM, page 852)

As long as we can decipher the neocortical algorithm and come up with a science of patterns, we can apply it to any system that we want to make intelligent. And one of the great features of neocortically inspired circuitry is that we won’t need to be especially clever in programming it. (Friday, January 09, 2015, 06:46 PM, page 922-24)

All our knowledge of the world is a model based on patterns. (Friday, January 09, 2015, 06:49 PM, page 932)

neurons are slow, so in that half a second, the information entering your brain can only traverse a chain one hundred neurons long. That is, the brain “computes” solutions to problems like this in one hundred steps or fewer, regardless of how many total neurons might be involved. (Friday, January 09, 2015, 08:05 PM, page 968-70)

In fact, it’s almost impossible to think of anything complex that isn’t a series of events or thoughts. (Friday, January 09, 2015, 08:14 PM, page 1026-27)

Truly random thoughts don’t exist. Memory recall almost always follows a pathway of association. (Friday, January 09, 2015, 08:17 PM, page 1040-41)

At any point in time we recall only a tiny fraction of what we know. (Friday, January 09, 2015, 08:20 PM, page 1068)

• The neocortex stores sequences of patterns. • The neocortex recalls patterns auto-associatively. • The neocortex stores patterns in an invariant form. • The neocortex stores patterns in a hierarchy. (Friday, January 09, 2015, 08:24 PM, page 1016-19)

Your ability to easily recognize the song in any key indicates that your brain has stored it in this pitch-invariant form. (Friday, January 09, 2015, 08:43 PM, page 1187-88)

Prediction is so pervasive that what we “perceive”—that is, how the world appears to us—does not come solely from our senses. What we perceive is a combination of what we sense and of our brains’ memory-derived predictions. (Monday, January 12, 2015, 10:11 AM, page 1260-62)

Prediction is not just one of the things your brain does. It is the primary function of the neocortex, and the foundation of intelligence. (Monday, January 12, 2015, 08:55 PM, page 1295)

When you listen to a familiar melody, you hear the next note in your head before it occurs. (Monday, January 12, 2015, 08:59 PM, page 1326-27)

When that prediction is wrong, your attention is immediately aroused. This is why we have difficulty not looking at people with deformities. (Monday, January 12, 2015, 09:13 PM, page 1381-82)

Prediction, not behavior, is the proof of intelligence. (Tuesday, January 13, 2015, 09:26 AM, page 1520)

Yet we know the brain solves it in a few steps, so the answer can’t be that difficult. (Wednesday, January 14, 2015, 08:02 AM, page 1621-22)

Things I see lead to precise predictions about things I will feel and hear, and the other way around. (Wednesday, January 14, 2015, 06:07 PM, page 1691-92)

As Mountcastle pointed out, the motor cortex looks like the sensory cortex. (Wednesday, January 14, 2015, 06:10 PM, page 1718-19)

The brain treats abstract and concrete objects in the same way. They are both just sequences of patterns that occur together over time in a predictable fashion. (Thursday, January 15, 2015, 09:14 AM, page 1830-31)

The odds of numerous input patterns occurring in the same relation over and over again by sheer coincidence are vanishingly small. A predictable sequence of patterns must be part of a larger object that really exists. (Thursday, January 15, 2015, 09:15 AM, page 1835-37)

the human cortex has an estimated several hundred million microcolumns. (Monday, January 19, 2015, 06:41 PM, page 1999)

Thus the information in layer 1 represents both the name of a sequence and the last item in the sequence. In this way, a particular column can be shared among many different sequences without getting confused. Columns learn to fire in the right context and in the correct order. (Monday, January 19, 2015, 11:15 PM, page 2115-17)

The higher the unexpected pattern needs to go, the more regions of the cortex get involved in resolving the unexpected input. (Thursday, January 22, 2015, 09:50 AM, page 2250-51)

Creativity can be defined simply as making predictions by analogy, something that occurs everywhere in cortex and something you do continually while awake. (Monday, January 26, 2015, 11:09 PM, page 2578-79)

In fact, highly creative works of art are appreciated because they violate our predictions. (Monday, January 26, 2015, 11:15 PM, page 2626-27)

An expert is someone who through practice and repeated exposure can recognize patterns that are more subtle than can be recognized by a nonexpert, such as the shape of a fin on a late-fifties car or the size of a spot on a seagull’s beak. (Monday, January 26, 2015, 11:19 PM, page 2653-55)

Our brains are always looking at patterns and making analogies. If correct correlations cannot be found, the brain is more than happy to accept false ones. Pseudoscience, bigotry, faith, and intolerance are often rooted in false analogy. (Tuesday, January 27, 2015, 10:53 AM, page 2725-27)

I believe consciousness is simply what it feels like to have a neocortex. (Tuesday, January 27, 2015, 10:59 AM, page 2764)

To the cortex, our bodies are just part of the external world. (Wednesday, January 28, 2015, 07:09 PM, page 2820)

The cortex builds a model of your body but it can’t build a model of the brain itself. Your thoughts, which are located in the brain, are physically separate from the body and the rest of the world. Mind is independent of body, but not of brain. (Wednesday, January 28, 2015, 07:10 PM, page 2824-26)

Imagining requires a neural mechanism for turning a prediction into an input. (Thursday, January 29, 2015, 09:20 AM, page 2846)

Most of what you perceive is not coming through your senses; it is generated by your internal memory model. (Thursday, January 29, 2015, 09:22 AM, page 2860-61)

Moral reasoning, both the good and the bad, is learned. (Thursday, January 29, 2015, 09:24 AM, page 2877-78)

There will be no need or opportunity for anyone to program in the rules of the world, databases, facts, or any of the high-level concepts that are the bane of artificial intelligence. (Friday, January 30, 2015, 11:02 AM, page 2948-49)

Intelligence is measured by the predictive ability of a hierarchical memory, not by humanlike behavior. (Friday, January 30, 2015, 11:03 AM, page 2959)

Self-replication does not require intelligence, and intelligence does not require self-replication. (Sunday, February 01, 2015, 03:05 PM, page 3034-35)

the Turing Test, by equating intelligence with human behavior, limited our vision of what is possible. By first understanding what intelligence is, we can build intelligent machines that are far more valuable than merely replicating human behavior. (Wednesday, February 04, 2015, 11:59 AM, page 3276-78)

The set of cells remains active as long as events that are members of the sequence are occurring (e.g., a set of cells that remains active as long as any note in a melody is being heard). (Wednesday, February 04, 2015, 01:56 PM, page 3365-66)