A Brief History of Artificial Intelligence - Michael Wooldridge

 Sub-title: What It Is, Where We Are, and Where We Are Going

I've pursued learning about artificial intelligence this year as AI technology is evolving in leaps and bounds.  This book was written a few years ago so it provides a look in the rear view mirror.  

It took me some months to read the entire book.  In the end, my thoughts were "so what?"  It's an example of how quickly information becomes out of date these days.  I did find the references listed at the end of the book to be useful in providing other historical sources.  

Quotes:

"social media platforms have an incentive not to show you the truth but to show you what you like. The role of AI in this is to figure out what your preferences are based on what you like..you aren't going to get a balanced picture." 

Published: 2020  Read: December 2024  Genre: Science

Co-Intelligence - Ethan Mollick

 Sub-title: Living and Working with AI

I've been pursuing learning about artificial intelligence since August of 2023, have taken classes and listened to webinars and watched dozens of YouTube videos.  ChatGPT is now part of my daily work, answering questions, explaining concepts, and assisting in all sorts of tasks.  I pre-ordered this book when I discovered the author's newsletter One Useful Thing.  There is so much to digest and ponder in this new field.  Highly recommended.

p. 35 skewed view of the world because data is chosen by mostly male computer scientists

Reinforcement Learning from Human Feedback (RLHF) the concept of humans reviewing and adjusting AI output; not foolproof.  Can use "prompt injection" i.e., giving the AI instructions via prompts and "jailbreak" AIs i.e., having it go against its set limits.

In 1966 a program called ELIZA simulated a psychotherapist.  It showed that creating an illusion of intelligence was possible by using simple tricks and exploiting the human tendency to project meaning and emotions onto machines.

The ability to predict what others are thinking is called theory of mind, thought to be exclusive to humans.

AI relies on patterns for answers, from learned data, not memory.  Anything that requires exact recall is likely to create a hallucination.

LLMs are connection machines.  They can connect seemingly disparate items in novel and creative ways. A test for creativity is called the Alternative Use Test (AUT). AIs excel at this.

"Sparks of AI General Intelligence" March 2023 - Eric Horvitz

Published: 2024  Read: April 2024  Genre: Science

The World I See - Dr. Fei-Fei Li

 Sub-title: Curiosity, Exploration, and Discovery at the Dawn of AI

This was a wonderful read that sheds a bright light on the early innovations in AI and one of its fascinating developers.  Dr Li weaves the story of her immigration and experiences in the US with the evolution of the technology.  She beautifully illustrates the power of family, mentors, and self-determination on realizing the intellect of a brilliant woman.  Highly recommended.

Quotes/Notes:

[on her mentor helping her and learning about her culture]..he read Chinese classics like Dream of the Red Chamber, Romance of the Three Kingdoms, and Journey to the West.

"But the failures are not in vain.  Mistakes trigger corrective signals, rippling across the network's tens of millions of constituent parts, each assessed for its contribution to the results and pushed, proportionately, to behave differently next time.  It's the simplest form of learning--to do less of whatever failed, and more of whatever didn't."

[on psychologist Eleanor Rosch and categorization and creating a vocabulary for understanding hierarchy - made me think of looking at DNA this way] "..superordinates and subordinate categories ..when taken as a whole, such hierarchies look like trees.  Moving toward the root means less specificity and differentiation, while moving toward the leaves, the farthest ends of each branch means more. "

[Anatol Holt] "AI was a technology that can make the perfect chess move while the room is on fire".

[on Jeremy Wolfe, cognitive scientist - made me think of doing a study on how blind and deaf perceive the world] "...concept of 'gist' ...directly confronted the question of what, exactly, humans perceive when they glance at the real world"  

"adversarial attacks" -  feeding information to directly confuse the AI

"explainability" or "explainable AI" - reducing neural networks' almost magical deliberations into a form humans can understand.

Published: 2023 Read: January 2024 Genre: Science