People who have influenced my thinking on AGI and the mind

There are a lot of great people who continue to influence my thinking about how the mind works. Go read a bunch of their work if you can.

Joshua Tenenbaum

Scholar profile: https://scholar.google.com/citations?hl=en&user=rRJ9wTJMUB8C

Joshua is a powerhouse and continues to do cutting edge research on how to build real AI. He does breadth and depth research into cognition, innate intelligence, intuitive physics, and grounded learning.

He has worked on probabilistic programming languages, which is an attempt to make a programming language that acts more like a human mind with probabilistic models.

On top of that he has done lots of research integrating his ideas into current neural network research.

He got me thinking about how babies build models of our world and how they learn intuitive physics. He is one of the most successful people I have ever seen combine theory with engineering.

Some papers I like:

Building Machines That Learn and Think Like People: https://arxiv.org/abs/1604.00289

Douglas Hofstadter

Most people in computer science know him from his book “Gödel, Escher, Bach: an Eternal Golden Braid”. It is a hard read. I actually recommend his other books as a clearer understanding of the mind.

“I am a Strange Loop”

“Surfaces and Essences”

“The Mind’s I”

He believes analogy and categorization are the main ingredients for human level intelligence. My interpretation of his interpretation is that the mind is composed of autonomous interlinked models (categories). The human mind can use these models/categories/concepts and manipulate them with analogies. To be able to make analogies, emotion must be involved to be able to choose “what feels right”. I don’t think a computer will ever be able to make analogies like us unless computers can have some form of emotion. Interestingly, Douglas has never published a research paper, only wonderful books. Melanie Mitchell, his student, is considered his successor and publishes papers.

Danielle Dannet

I have not read many of his papers, but I have some of his books and articles about consciousness.

https://scholar.google.com/citations?hl=en&user=3FWe5OQAAAAJ

One of my favorite books from him is “Intuition Pumps And Other Tools for Thinking”.

He has a theory called “Multiple draft models” where he states: “All varieties of perception—indeed all varieties of thought or mental activity—are accomplished in the brain by parallel, multitrack processes of interpretation and elaboration of sensory inputs. Information entering the nervous system is under continuous ‘editorial revision.'”

He has taught my how philosophers uses their tools such as “intuition pumps” to reason about problems. For example, take some idea you are thinking about and exaggerate it in your mind’s simulation and see what happens. Does the idea still hold?

Jeff Hawkins

His book “On Intelligence” is the book that pushed me to believe that AGI could potentially be discovered in our lifetime.

Jeff Hawkins is one of the only researches I know who is actively focused on learning from the brain to create actual intelligent computer systems. He was also a successful entrepreneur twice in Silicon Valley, and so has the money and knowledge to mobilize teams of people. His basic premise is that there is one main learning algorithm in the Neocortex (proposed by Vernon Mountcastle) and that we must look to human brains to discover intelligence. And so his research is focused on extracting this neocortical algorithm into a computer algorithm. As of late, he has been focusing his work on grid cells.

Helen Keller

At the age of 18 months, Helen Keller lost her hearing and vision rendering her blind and deaf for the rest of her life. Despite that, she was still able to be an intelligent and productive member of society. She interacted through the world primarily with touch. She wrote a wonderful book talking about how she perceives the world. She even describes concepts that can only be seen such as colors. She showed me several things: The core of intelligence does not rely on vision, the intelligence algorithm is the same across different modalities, and that the combination of sensors and motion are required for intelligence. I highly recommend you read her book “The World I Live In”.

Her definition of duty: “I feel as if I were going forward in a straight line, bound to arrive somewhere, or go on forever without swerving to the right or left”.

From her writing: “Touch cannot bridge distance,—it is fit only for the contact of surfaces,—but thought leaps the chasm. For this reason I am able to use words descriptive of objects distant from my senses.”

Another quote from her: “When I learned the meaning of “I” and “me” and found that I was something, I began to think. Then consciousness first existed for me. Thus it was not the sense of touch that brought me knowledge. It was the awakening of my soul that first rendered my senses their value, their cognizance of objects, names, qualities, and properties.”

One question I still have: If Helen’s vision system did not develop until 18 months, would she still be intelligent like us? In other words, if she was truly born blind and death and those neural structures did not develop, would she still have the capacity to have human level intelligence?

Stanislas Dehaene

He is a researcher focused on cognitive neuroscience. He introduced to me the idea that we can measure certain aspects of consciousness in an empirical way. For examples many of his studies uses images that are flashed onto the human retina for a time period below the threshold that would make the subject consciously aware of the stimulus. He also introduced to me the idea of Global Workspace Theory.

https://scholar.google.com/citations?hl=en&user=2Dd5uoIAAAAJ

Karl Friston

Considered the most prolific neuroscience researcher alive. His most famous idea is the free-energy principle which suggests that all living things, including the human mind is minimizing for surprise. This is interesting for machine learning because this idea could be used to figure out how to build true unsupervised learning systems versus our current systems. He also introduced me to Markov Blankets (although that idea is from Judea Pearl).

https://scholar.google.com/citations?hl=en&user=q_4u0aoAAAAJ

The Free-energy principle: A Unified brain theory?

Leave a comment

Your email address will not be published.