Sign in with Facebook
Nora's Reviews > The Alignment Problem: Machine Learning and Human Values
The Alignment Problem: Machine Learning and Human Values
by
by
4/5
FINALLY finished this. I actually started in June 2022 and abandoned it. For someone with no knowledge of ML, this book is probably too dense and convoluted to get anything out of. There's a lot in there. After getting a bit more exposed and coming back to it, I was able to get all the way through. I get that Brian wanted to emphasize the different researchers and philosophers he interviewed, but the constant name dropping didn't do the book much good. Like, just tell me what they did. BUT this is also one of the books I have quoted the most in conversations about tech. The chapter on Fairness - the COMPAS model and the use of AI in criminal justice - was miles more interesting than the other ones. When fairness is mathematically impossible, where do we turn? This book is also pretty good if you want a background of the field of ML, not the pure technical stuff. Lots of history and connections to philosophy, neuroscience, psychology, etc. Just pick and choose your favorite ideas out of the million that are briefly mentioned in here and go down your own rabbit hole.
FINALLY finished this. I actually started in June 2022 and abandoned it. For someone with no knowledge of ML, this book is probably too dense and convoluted to get anything out of. There's a lot in there. After getting a bit more exposed and coming back to it, I was able to get all the way through. I get that Brian wanted to emphasize the different researchers and philosophers he interviewed, but the constant name dropping didn't do the book much good. Like, just tell me what they did. BUT this is also one of the books I have quoted the most in conversations about tech. The chapter on Fairness - the COMPAS model and the use of AI in criminal justice - was miles more interesting than the other ones. When fairness is mathematically impossible, where do we turn? This book is also pretty good if you want a background of the field of ML, not the pure technical stuff. Lots of history and connections to philosophy, neuroscience, psychology, etc. Just pick and choose your favorite ideas out of the million that are briefly mentioned in here and go down your own rabbit hole.
Sign into Goodreads to see if any of your friends have read
The Alignment Problem.
Sign In »
Reading Progress
August 3, 2023
–
Started Reading
August 3, 2023
– Shelved
August 7, 2023
–
Finished Reading
